STRATEGIES AND ACHIEVEMENTS IN FRUIT BREEDING AND SELECTION AT MARDI

ABSTRACT
The fruit collection and germplasm centres for fruits were established at MARDI to provide breeders with the basic resource, i. e., genetic variation for crop improvement. The accessions in the fruit collection consists of varieties, cultivars, clones or advanced breeding lines and was started in 1972 at Serdang. The germplasm, whose accessions were collected at random and restricted only to indigenous fruit species, was started in 1982 at Kemaman. This was a joint effort with the IBPGR.

Characterization and evaluation, important in providing an inventory of the genetic variation, were completed for papaya and banana and partially completed for perennial fruits like durian, mango and cashew nut.

In breeding and selection of clonally propagated fruit species, the basic step is to identify and select the 'winners' and subsequently testing them over time and space to establish the repeatability (stability) of their good performance. These 'winners' are, however, time-related because of the changing consumer preference and industry demands. They will form the base for further improvement of the varieties in terms of increasing yield, developing pest and disease resistance and other requirements demanded by the industry from time to time. For clonally propagated fruits, selection of rootstocks and increasing productivity of durian clonal populations using mixtures of compatible genotypes are also given consideration.

For seed-propagated fruit species, the breeding and selection methodology for improving open pollinated species (watermelon) and self-pollinated species (papaya) were included.

The achievements of the breeding programme on the whole were discussed in the light of the development of Nanas Johor pineapple, the MARDI Backcross Solo papaya, new accessions of Mas and Rastali bananas, the C11 cashew nut clone, and recommendation of mixtures of durian clones based on compatibility and xenic effects of the pollination.

INTRODUCTION
Fruit breeding and selection before the advent of MARDI was carried out by the Department of Agriculture and in the case of pineapple, by the Malayan Pineapple Industry Board (MPIB). The Department of Agriculture placed emphasis on selection of promising clones based usually on single tree performance and kept a register of these selections with a brief description on quality and the date and place they were selected from (see ANON, 1980). Some of the entries dated as far back as 1933, and several of them, for example, D2, D24 and R3 which were among the earliest registered, had continued to be popular clones till the present day. For pineapple, the notable success scored by the MPIB was the selection of 'Masmerah' from a population of Singapore Spanish (WEE, 1974).

For seed-propagated fruit types like papaya and melons and those that are of lesser economic importance such as cempedak, nangka, ciku, duku langsat and mangosteen, breeding work had been fairly neglected in the pre-MARDI era.

Research on fruits began to follow a more systematic approach with the formation of the Fruit Unit in MARDI in 1972, and six years later, it was promoted to Branch level. It subsequently graduated to Divisional status in 1984. Research on breeding and selection for high yield and good quality fruit varieties is one of the seven programmes under the present structure of the Fruit Research Division. This paper discusses the strategies adopted in the Programme for development of fruit varieties using different methodologies and highlights some of the achievements made.

CONSERVATION AND CHARACTERIZATION OF GENETIC RESOURCES
This is one of the two sub-programmes for development of fruit varieties, the other being breeding and selection (Figure 1).

Chart of development program.

The fruit collection and germplasm for fruits were established at MARDI to provide breeders with the genetic variation necessary for crop improvement. The two terminologies, collection and germplasm, both have a similar function, i. e. assembling a range of accessions of the fruit species to provide the genetic variation vital for the breeding programme. The difference, however, lies in the nature of the accessions and the mode of collection. In fruit collection, the accessions are usually varieties, cultivars, clones or advanced breeding lines, the collection of which is non-random. On the other hand, accessions in germplasms are obtained at random from natural populations and in this respect, a representation of the genepool is emphasized.

The fruit germplasm was initiated in 1982 in collaboration with IBPGR, and only indigenous species of economic importance such as durian, rambutan, banana, cempedak and mango were given emphasis. The sampling was stratified according to the occurrence of the fruit type in the localities, and covered fruit areas in Peninsular Malaysia, Sabah, Sarawak and Brunei. The number of accessions collected for the different fruit species is presented in Table 1. All the materials were established at MARDI, Kemaman.

Table 1. Number of accessions by fruit types in the collection and germplasm in MARDI

Fruit TypesCollectionGermplasm
(MARDI-IBPGR)

Durio88586
Nephelium62191
Artocarpus3838
Mangifera111259
Musa14817
Lansium587
Garcinia-21
Ananas75-
Achras14-
Citrus94-
Papaya52-
Psidium17-
Averrhoa16-
Annona15-
Rare Fruits-46

The fruit collection centre in Serdang for durian, mango, banana, papaya, cempedak, nangka, ciku, guava, starfruit and custard apple was initiated in 1972. For other fruit types, the collection centres were distributed over several MARDI stations, e.g., rambutan, cempedak (Bukil Ridan) , mango (partial duplication in Kluang), pineapple (Pontian) and citrus (Cameron Highlands, Jerangau). At the start, the collection at MARDI, Serdang consisted of the duplication of the clonal materials registered by the Department of Agriculture. More accessions were added to the collection from time to time, obtained from farmers' fields, fruit exhibitions, as well as from neighbouring countries. The number of accessions in the collection for each fruit species is presented in Table 1. The collection of non-indigenous fruit types, particularly chiku, guava and star fruit, were considerably smaller because of difficulties in introduction of these materials.

Characterization and evaluation of the fruit collection and germplasm permit the breeder to take stock of the variability accessible to him for manipulation in crop improvement. This important aspect in collection and germplasm work allows one to know what one has in order to expect what one can achieve through breeding efforts.

This aspect of work has been completed and documented for papaya (CHAN, 1985a) and banana (SITI HAWA JAMALUDDIN, 1986) and partially completed for durian, mango and cashew nut. The description of the collection for a particular character is based on the distribution pattern of the accessions, the mean, range and CV%. In the description of the papaya collection for total soluble solids %, for example, it was found that the distribution of the accessions was normal, with a range of 9.30% to 16.67%, a mean of 12.64% and a CV of 12.32%. (CHAN, 1985a) (Figure 2). With such information, the breeder will be aware of the behaviour of the population and the limits of variation within which he can exploit.

BREEDING and SELECTION
This is the second sub-programme and the treatment of this topic will be divided into two parts, i.e., clonally propagated and seed propagated species (Figure 1).

Clonally propagated fruit species
Many of the woody perennial fruit species can be clonally propagated by grafting, and to a lesser extent by air-layering and cuttings. Besides reducing the period of juvenility, clonal propagation allows the fixation of heterozygosity, giving rise to homogeneous populations within which the individuals are heterozygous in genetic make-up.

Improvement of such fruit species, therefore, appears uncomplicated. The primary aim is to select for the 'winner'; very often this being a single tree of outstanding quality, and regenerating the genotype by vegetative means. For indigenous species, the sources where likely 'winners' may be found are from fruit collections, germplasm centres and farmers' fields (dusuns) where seedling populations represent a reservoir of genetic variation. For non-indigenous species, cultivars or varieties bred elsewhere may be likely 'winners' if they are well-adapted here.

The preliminary selection of the apparent 'winners' however, is often based on qualitative characters such as fruit appearance, flesh, colour, pulp thickness and flavour and quantitative characters like yield, and its stability cannot be elucidated from data obtained from a single maternal tree from which it was selected. Following the preliminary selection therefore, a genotype x environment trial is carried out to test the relative performance and stability of the selected genotypes in the company of standard cultivars over several locations and seasons.

Genotype X Environment trials
At this point in time, G x E trials have been initiated for durian, mango, rambutan, papaya, banana, pineapple, starfruit and chiku. The recently concluded trial on banana will be cited as an example of the methodology for establishing performance and stability of the genotypes.

The G x E for banana did not place emphasis on the difference in performance and stability between cultivars i.e. between Mas, Rastali, Embun or Berangan. Inter-cultivar differences are expected to occur, but more importantly, the futility of such comparisons is reflected by the strong socio-preferences of farmers for a particular cultivar. Farmers do not invariably change over to another cultivar simply because it is 'more stable' or better yielding.

G x E for banana is used for establishing genotypic differences and stability of accessions within a cultivar. The trial involving four Rastali accessions tested over four locations is presented in Figure 3. Two statistics are used for describing stability i.e. the slope (b) and the deviation (SE) from the linear regression. The results showed that the accessions from Kuala Kangsar were higher yielding and consistently so over all locations compared with the others. The good stability of this accession, as well as 'Pontian', are shown by the slope of the regression nearly equal to one as well as the small deviations from the regression (Figure 3). These two are current recommendations of Rastali banana.

Further improvement of varieties
After the long-drawn process of selecting the 'winners' and proving their stability and performance in G X E trials, some work may still be necessary to improve the varieties further. 'Winners', it must be pointed out here, are 'time-related' because consumer preferences and industry demands often change with time and the breeding programme must be dynamic enough to meet such changes. Varieties recommended following G x E trials can form the base for choosing parents in crosses to improve their deficient characters.

In the pineapple industry, Masmerah and Gandul (Singapore Spanish) are varieties that are good examples of such 'base varieties' which have been around for at least two decades and which require changes. Their golden flesh is a definite asset for canning of premium priced fancy cuts, but their yield is only 40% to 60% that of Cayenne, the standard variety of other pineapple-growing countries.

Hybridization between the Singapore Spanish and the large-fruited Cayenne, and selection in the segregating F1, produced a new variety, the Nanas Johor (CHAN and LEE, 1985). The new hybrid had better vigour, yield and fruit quality compared with its commercial contemporaries.

Nanas Johor, which showed considerable yield improvement, still has some shortcomings. It has occasional incidence of marbling fruit and its core is thick, particularly in larger fruits. A backcross to Singapore Spanish (resistant to marbling) is now in progress to further improve this new hybrid.

For other fruit types, some 'base' varieties have been identified following selection and clonal testing, but the majority of them have some weaknesses which should be redressed to elevate the worth of these varieties. C11 cashew clone, for example, is the best-yielding clone developed in this country so far. The results of the trial of 16 selected clones at MARDI showed that C11 was at least 30% better than the next best, i.e., C21 and about two to six times better than the rest (SALLEH HASSAN, CHAI and BAKRIMAT LASA, 1986). Its mean nut weight of 5.6 g, however, is rather low and this should be improved. For mangoes, Harumanis is an accepted variety for commercial cultivation, but it has occasional problems with 'soft-nose' of the fruits. D24, a long-standing popular durian clone, is rather sensitive (less hardy) during the establishment stage compared with the majority of Thai clones. All these examples shows that while 'base' varieties exist for a number of fruit species, further breeding work is required to improve their excellence.

Increasing productivity of clonal populations
Another important aspect worth consideration is increasing the productivity of the clonal population as a whole. The present density for the majority of perennial woody fruit species ranges from 70 to 90 trees per hectare. Selection of dwarfing rootstocks to induce smaller plant structures and develop high-density efficient cropping systems should receive more attention. With the attendant advantages of precocity and the convenience in fruit harvest, the productivity per unit area per unit time will be increased. The East Malling dwarfing apple rootstocks and the success of the meadow and hedge row orchards should be emulated.

For the self-incompatible local durian clones, breeding work should not be confined to selection of a single clone because recommendation of 11 per se as an orchard monocrop will be disastrous. In this respect, complementary pollinator clones compatible with, and which have favourable xenic effects on the recommended genotype, must be selected as well.

Compatibility studies on durian established that D10 was highly compatible with the recommended D24 and fruit development was better, yielding larger and more rounded fruits with fewer empty locules (SHAARI, ZAINAL ABIDIN and MOHD. SHAMSUDIN, 1985). Recent crosses of D24 with D99 indicated that the latter which is self-compatible, is another potential pollinator for D24. Fruit set was in the region of 60% and favourable xenic effects were exhibited because pulp colour appeared richer with a smaller number of seeds compared with fruits pollinated by D10. Further, the number of shrivelled seeds (an important quality consideration) was considerably higher (SHAARI, pers comm.).

The clonal composition presently recommended is 60% D24, 30% D99 and 10% D10 or D98.

Seed propagated fruit species
The methodology and approach in breeding and improvement of seed propagated fruit species depend on the breeding systems of the species i.e. whether they are open-pollinated or self-pollinated.

Open-pollinated fruit species: Watermelon
Open-pollinated varieties were introduced for testing at Sungai Baging and the results indicated that Crimson Sweet was outstanding in quality and yield (ANON, 19B4).

For production of seeds of the Crimson Sweet, pollen of the variety was collected and bulked from random plants in the population and sib mated with the female flowers. The seeds developed in the fruits were bulked and tested with the original Crimson Sweet to determine if there was any genetic drift. Preliminary results indicate that this methodology was effective in producing seeds that have similar performance as that from the original source (SALLEH HASSAN, pers. comm.).

In the project for production of F1 seeds, three lines of Sugar Lee, nine Crimson Sweet, two Leesburg and one each of Charleston Gray, Hawkesbury and Cathourn Gray were selfed to generate inbred lines. Some inbreeding depression in terms of plant vigour and seed fertility were evident. The selfing will be carried to the F7 generation, after which the inbred lines will be selectively tested for their combining abilities in diallel crosses for choosing parents for production of F1 watermelon seeds.

Self-pollinated fruit species: papaya
Several varieties of papaya, notably the Hawaiian Solo lines and a few local accessions viz. Subang 6, Sitiawan and Batu Arang, are basically self-pollinating in hermaphrodite flowers because of cleistogamy (anthers dehisce before flowers open).

In development of varieties for processing as well as for dessert fruits for local consumers, large fruit size, high yield and sweet, firm, red flesh are important criteria. As a start, eleven populations collected from various geographical locations were tested and the majority were found to be very variable in terms of fruit weight and total soluble solids % (CHAN and 001, 1975). Selection of the promising progenies and continuously selfing them for several generations considerably reduced the variation and improved yield, particularly in the case of Sitiawan (CHAN, 1985b).

The backcross programme was used to develop varieties with quality fruits for local as well as for the export markets. The recurrent parent was the Sunrise Solo which has excellent fruit qualities, but its small fruit size (0.3 - 0.4 kg) obtained under local conditions, was unsatisfactory (CHAN and TEE, 1975) . The non-recurrent parent was Subang 6, a fairly stable, local variety bearing large fruits (1.5 kg) with red, firm flesh. For improvement of fruit size of the Solo (a quantitative character) an F2 generation was produced after each backcross to Solo to provide greater precision in selection for size in the segregating F2 population.

Presently, two lines of B2F5, i.e., Line 19 and Line 20 are recommended for cultivation. The uniform fruit size around 600 to 700 g is about twice that of the original Solo and the total soluble solids %, in the region of 13% to 14%, is very close to that of Sunrise Solo. G x E testing over several environments including a peat location, indicated that the Backcross Solo was very stable in yield and total soluble solids %.

Results of the market testing of the Backcross Solo fruits carried out at the FIMA-RANTEI Restaurant in Kuala Lumpur were very encouraging, with the majority of the respondents receiving the fruits favourably (CHAN, 1984).

Breeder stock seeds have been recently supplied to the Department of Agriculture for production of foundation and certified seeds to growers. In the meantime, MARDI had entertained requests from both public and private agencies and growers in the supply of planting materials of this new papaya variety.

TSS% classesYield response of 4 Rastali accessions over 4 locations
Figure 2. Distribution of TSS% for the papaya collectionFigure 3. Yield response of 4 Rastali accessions over 4 locations

ACKNOWLEDGMENT
The author thanks the Director-General of MARDI for permitting this paper to be presented in the Symposium.

REFERENCES
ANON (1980). Registered fruit clones of the Department of Agriculture. Cawangan Pengeluaran Tanaman, Jabatan Pertanian Malaysia. 49 pp.

ANON (1984). Lapuran Tahunan 1984. Bahagian Penyelidikan Buah-Buahan, MARDI, Serdang, Selangor, 143 pp.

CHAN, Y.K. (1984). A Report on the market testing and promotion of the new MARDI Backcross Solo papaya. MARDI Fruit Branch info. leaflet. no. 16. 11 pp.

CHAN Y.K. (1985a) Maintenance, characterization and utilization of the papaya genetic collection in Malaysia. Paper presented at the SABRAO V International Congress, Bangkok. 25-29 Nov. '85.

CHAN, Y.K. (1985b). Evaluation of the performance and stability of papaya varieties bred at MARDI. MARDI Res. Bull. 13:1. 1-7.

CHAN, Y.K. and S.C. 001 (1975). Preliminary studies on papaya selection in Malaysia. Malays. agric. J. 50, 67-77.

CHAN, Y.K. and T.S. Tee (1975). Studies on Sunrise Solo papaya in Malaysia. MARDI Report no. 36. 10 pp.

CHAN, Y.K. and C.K. LEE. (1985). The Hybrid 1 pineapple: A new canning variety developed at MARDI. Teknologi Buah-buahan, 1:1, 24-30.

SALLEH HASAN, T.B. CHAI and BAKRI MAT LASA. (1986). Evaluating yield performance and stability of 16 clones of cashew. MARDI Res. Bull. (Submitted for publication).

SHAARI, A.R., M. ZAINAL ABIDIN and O. MOHD. SHAMSUDIN, (1985). Some aspects of pollination and fruit set in durian, Durio zibethinus Murr., cultivar D24. Teknologi Buah-buahan, 1:1, 1-4.

SITI HAWA JAMALUDDIN. (1986). Characterization, evaluation and utilization of the banana germplasm in Malaysia. Paper presented at the Simposium Buah-Buahan Kebangsaan, Serdang, Selangor, 8- 9 April 1986.

WEE, Y. C. (1974). The Masmerah: A new cultivar for the Malaysian pineapple industry. Wld. Crops. 26, 64-67.

Y.K. CHAN
Fruit Research Division MARDI, Serdang
Article from Proceedings of the National Fruit Symposium 1986

DATE: March 1990

* * * * * * * * * * * * *