EVALUATION AND BREEDING USE OF SUGARCANE GERMPLASM B.T. Roach, CSR Field Experiment Station, Macknade, Queensland, Australia.

Transcription

1 EVALUATION AND BREEDING USE OF SUGARCANE GERMPLASM B.T. Roach, Breeding CSR Field Experiment Station, Macknade, Queensland, Australia ABSTRACT This paper summarises data collected during a three-year survey of 560 clones of species and genera of the Saccharum complex maintained at Macknade. Using six quantitative characters of economic importance and within-species grouping on geographic origin, genetic differences in S. spontaneum are shown to be associated with geographic origin. The situation in S. robustum is less clear and there is an apparent uniformity within Erianthus arundinaceus, irrespective of geographical origin. The importance of genetic divergence of parents in maximising heterosis is discussed. It is suggested that the north-east Indian group of S. spontaneum merit sampling in breeding, as do the south-east Asian, Sulawesi and Kalimantan groups. Sampling of the remaining geographical groups of S. spontaneum is also merited, together with the 5 groups of S. robustum delineated on cytological and chemotaxonomic bases. To rapidly utilise selected germplasm from a wide range of sources, a program of bridging crosses is suggested as a population improvement program outside the main breeding program. INTRODUCTION The use of wild germplasm in breeding crop plants is a relatively recent phenomenon (Hawkes18) and sugarcane has been regarded as one of the crops benefitting most from early use of wild relatives (Buddenhagens and Stalker43). However some'40 years after the initial and most spectacular hybridisation successes were achieved, there was concern at the slow rate of breeding progress and the narrow genetic base of modern hybrid sugarcanes (Arceneaux3 and Danielsl0). Nobilisation or "base broadening" programs were commenced with Saccharum spontaneum in Australia (Roa~h~~l~~), Barbados (Walker45,48) and Louisiana (Dunckleman15), while in Hawaii there had been continued interest in the collection and use in crosses of both S. spontaneurn and S. robusturn (Heinz19). More recently, less closely related Miscanthus and Erianthus have been used in Taiwan (Chen9) and Barbados (Walkerso) respectively. The need to introduce wild germplasm into a breeding population is debatable. R~bbelen~~ has questioned whether the top productive varieties of the main crops Keywords: Wild germplasm, germplasm evaluation, genetic divergence, heterosis, bridging crosses

2 B.T. ROACH 493 today can still take advantage of the genetic diversity present in natural resources or whether they have passed the "point of no return" at which further addition of foreign variation does nothing but spoil the achieved high level of performance. This may be the situation in sugarcane (IIeinzZ0 and H ~garth~~~~~), although it is generally conceded that the initial hybridisations greatly increased the stress tolerance and geographical range of commercial sugarcanes. There has been no shortage of material for those wishing to increase the genetic diversity of their breeding populations. There have 'been numerous organised collections of representatives of sugarcane species and related genera and, whilst there has been severe erosion of this material in collections (Berding6), much still remains (Williams5 I). A sugarcane breeder contemplating use of new germplasm in a breeding program needs to know something of the characteristics of the many classes and clones of material available and how he might make a logical selection of parents from it. The objective of this paper is to provide some information on these aspects with data obtained from clones maintained in the germplasm collection at Macknade Experiment Station. MATERIALS AND METHODS The Macknade germplasm collection contains 560 clones, about 30% of the number in each of the two World Collections. It is reasonably representative of material deposited in these collections up to and including the 1976 ISSCT collection from Indonesia. Clones of S. spontaneum are maintained in duplicate in containers of 100 L capacity, other taxa being maintained in the field. Data were collected on 25 characteristics over a 3-year period ( ). The characteristics used and the rating systems adopted followed the IBPGR guidelines agreed at the 1981 meeting of the IBPGR working group on genetic resources of sugarcane (An~n.~). A minimum of 10 stalks per plot were taken for analyses. More than 10 stalks were required for thinner stalked clones, particularly those of S. spovtaneum, to provide the minimum 1 kg sample for processing. Samples were processed in a Jeffco cutter grinder. Juice was extracted from a 500 g sample of fibrated material in a Carver laboratory press, using a pressure of kpa for one minute. Expressed juice was analysed for brix, pol and conductivity. Residual fibre was dried at 105 OC to estimate fibre per cent fresh weight following correction for brix. RESULTS Data for 6 of the 25 characteis recorded on the collection are presented in this paper. The 6 were selected as they are based on quantitative measurement rather than visual rating. The results are summarised in Table I, I1 and 111. Data have been omitted for species and genera of little breeding interest or with few representatives, e.g. S. edule and Pennisetum sp. In the tables the grouping of clones on the basis of geographical origin is rather arbitrary, particularly in the case of India and south-east Asia. S. spontaneum clones of the Indian sub-continent have been grouped by state of origin into southern (Tamil Nadu, Karnataka, Kerala), central (Bihar and Orissa), north and north-west I I I I

3 TABLE I. Summarised data for leaf width and stalk thickness. Group No. of Leaf width - cm Stalk thickness - cm Clones Range Mean CV% Range Mean CV% S, spontaneum India - south central north & west north east South-east Asia, Sulawesi ' Kalimantan Other Indonesian Papua New Guinea Western Japan and Taiwan S. robustum lrian Jaya Other Indonesian Papua New Guinea R. arundinaceum lrian Jaya Kalimantan Sulawesi Other S. sinense S. barberi S. officinarum Named and bred (1) Indonesia (2) New Guinea (3) Other (4) Note : (1) All named and bred clones, irrespective of commercial use. (2) Includes IJ, IK, IM and IS clones from 1976 ISSCT collection. (3) Includes clones from the NG16 to the NG57 collections. (4) Clones from Fiji (14), Malaysia (2) and New Hebrides (2). (Rajasthan, Punjab, Nepal, Uttar Pradesh) and north-east (West Bengal, Assam, Manipur, Bangladesh). The south-east Asian group includes clones from Burma, Thailand, Vietnam and Malaysia, while the western group includes clones of African and Mediterranean origin, following Panje and B ab~~~. Leaf width The data in Table I show considerable regional variation within S. spontaneum. Of the Indian clones, the north-east group have distinctively wider leaves than those

4 Group No, of Brix Fibre O/O Fresh-Weight Clones Range Mean CVO/o Range Mean CVO/o S. spontaneum India - south central north & west north east South-east Asia Sulawesi Kalimantan Other Indonesian ' ~a"pua New Guinea Western Japan and Taiwan S. robustum lrian Jaya Other Indonesian Papua New Guinea R. arundinaceum lrian Jaya Kalimantan Sulawesi Other S. sinense S. barberi S. officinarum Named and bred (1) Indonesia (2) New Guinea (3) Other (4) of the other regions. Following a slight reduction in the south-east Asian region, leaf width again increases in the Indonesian region and markedly so in the Kalimantan group. The Kalimantan group differs quite significantly from the Sulawesi group and the distribution overlaps well into S. robustum and even into S. officinarum. The Papua New Guinea and Japan/Taiwan groups have fairly uniformly narrow leaves, while the wide range in the western group reflects its wide geographical spread. S. robustum shows wide variation in leaf width, particularly in New Guinea clones, with a distribution intermediate between S. spontaneum and S, officinarum. The S. sinense/s. barberi group also shows a wide distribution, intermediate between

5 BREEDING TABLE Ill. Summarised data for Sucrose O/O Juice and Ash % Juice. Group Sucrose % Juice Clones Range Mean CV% Ash % Juice S. spontaneum India - south - central - north & west - north east South-east Asia Sulawesi Kalimantan, Other Indonesian Papua New Guinea Western Japan and Taiwan S. robustum lrian Jaya Other lndonesian Papua New Guinea R. arundinaceum lrian Jaya Kalimantan Sulawesi Other S. sinense S. barberi ) S. officinarum l m e d and bred (1) Indonesia (2) New Guinea (3) Other (4) ; S. spontaneum and S. officinarum. Clones of Erianthus arundinaceus (sect. Ripidium Henrard) show very limited variation, irrespective of origin. Stalk thickness Data for stalk thickness (Table I) show a pattern similar to that for leaf width. The two characters are closely related, with a correlation of r = 0.95 for group means. Within S. spontaneum, the Kalimantan group has the highest proportion of thick stalked clones and a distribution extending well into S. robustum and even into S. officinarum. The very limited variation in E. arundinaceus is again apparent. Brix Compared to leaf width and stalk thickness, there is relatively little variation within S. spontaneum for this character (Table 11). Brix levels in E. arundinaceus are similar to those in S. spontaneum, with S. robustum showing a marked increase towards the levels found in S._sf'nense/S. barberi and S. officinarum.

6 The data in Table I1 show little variation within S. spontaneurn and similar levels in S, robusturn. Fibre levels in E. arundinaceus are uniformly higher than in other species, probably a reflection of the thick, hard rind of this species and generally dry, pithy nature of the internal stalk. S. sinense and S. barberi are the only groups with fibre levels extending into the S. officinarurn range. Sucrose % juice (pol) Means and distribution patterns in Table I11 are similar to those found for stalk thickness and leaf width. A correlation of 0.66 exists between leaf width and sucrose 070 juice for the means of the 11 S. spontaneurn regional groups. High sucrose forms of S. spontaneurn have been reported previously (Marim~tharnmal~~ and Venkatraman44) and in this study, relatively high levels are found in the north-east India and Kalimantan groups. E. arundinaceus occupies a transitional place between S. spontaneurn and S. robusturn. Only S. robustum overlaps into the distributions found in S. barberi/s. sinense and S. officinarurn. Conductivity ash % juice The use of conductivity as a measure of ash has been well established (MeadeZ7 and mull in^^^). Within S. spontaneurn, the lowest, or more "noble", levels of ash occur in the north-east India and Kalimantan groups, (Table 111). Levels in S. robusturn are generally below those in S. spontaneurn but uniformly high levels of ash occur in E. arundinaceus. Ash levels in S. sinense/s. barberi and S. dfficinarurn do not overlap to any degree with those of the wild species or genera. Group means show a negative correlation of between ash and sucrose % juice. DISCUSSION There have been numerous attempts to make a logical selection of wild germplasm prior to hybridisation. Selection for disease resistance, the original reason for using wild germplasm, has been most popular. Selection for improved smut resistance has been practised in the West Indies in S. spontaneum and its F1 progeny (Walker46) and in Taiwan in Miscanthus sinensis and M. floridulus (Lo25). Downy mildew resistance was also sought in the latter program. Clones of S. spontaneurn have been tested for cold tolerance (Irvine23924) and waterlogging tolerance (Roach38 and Srini~asas~~). Some of the disappointment in transfer of attributes of wild germplasm to commercial sugarcane hybrids may lie in the fact that phenotype of the wild donor clone is so vastly different from commercial hybrids. Leaf diseases, for instance, may be poorly expressed on narrow leaves of S. spontaneurn consisting of little more than a mid-rib. Some clones of S. spontaneurn showing high waterlogging tolerance, e.g. SES-528, have hollow stems (aerenchyma), hardly desirable in a commercial hybrid. There have been relatively few attempts to select wild germplasm on the basis of its yield components prior to hybridisation. Dunckelman and Breaux14 did take agronomic quality into account in conjunction with cold tolerance and smut

7 498 BREEDING resistance. Nair and Somarajan31 sought genetic resources for high early sucrose in a large sample of S. officinarurn, S. sinense and S. barberi. Roach36 found that selection within S. spontaneurn for percent sucrose prior to nobilisation would be effective in increasing sugar content of resultant F1 hybrids. Selection for cane yield would be only moderately effective. A lot of the hybridisation work carried out with sugarcane germplasm in the last two decades could best be described as general "base broadening". Walker4" noted the vagueness of this term and that the initial objectives of the West Indies program had been too broad. Stalker43 observed that genetic variability usually has little value in itself. General "base broadening" may be an acceptable objective in a large breeding program conducted on a population approach and with use of melting pot crosses, as outlined by Heinzlg. Otherwise sampling of available genetic diversity and incorporation of useful traits will be unacceptably slow. If phenotype is a poor guide to selection of basic germplasm for hybridisation and test crossing is a slow and costly way to exploit a gene pool, then what alternatives remain? The answer may lie in the observation by Stalker43 that "before a group can be fully exploited, the available plant materials and their relationships must be understood". The advantages of interspecific hybridisation in sugarcane, have been perhaps greater in the generation of hybrid vigour (heterosis) than in obtaining resistance to specific diseases. The importance of genetic divergence of parents in maximising heterosis is well e ~tablished~~~~~~~~~ and there have been some efforts to investigate this in sugarcane. Gill and Tripathi16 classified 30 varieties of Indian and foreign origin into 13 clusters on the basis of measurement of 12 characters. Nagatomi and Oshiro30 classified 52 clones of wild sugarcane germplasm by two methods of numerical taxonomy, cluster analyses and principal component analysis and found the former to be more useful. Their results were in agreement with those of M~rishima~~ that the pattern by cluster analysis corresponded to their geographical distribution. The study of Singh and Singh41, which found genetic differences to be unrelated to geographic origin, used 48 commercial varieties whose distribution is a recent man-made phenomenon. Some observations in sugarcane breeding programs support the concept that genetic diversity is related to hybrid vigour. At Macknade, large populations of F1 hybrids from crosses of Badila (S. officinarurn) with both the S. spontaneurn clones Tabongo (2n = 80, Celebes) and Mandalay (2n = 96, Burma) have been produced and evaluated. However a population of clones produced from crossing Fl's from each cross surpasses in vigour either of the F1 populations. Walker49 has also observed that hybrids between F1 clones from different S. spontaneurn sources can show heterosis for sugar content and yield. The sample of wild germplasm used to produce present-day hybrids (Roach35) may have been fortuitous in its diversity of origin and genetic constitution. The two principal S. spontaneurn sources were 2n = 64 from southern India and 2n = 112 from Java. Data in Table I to I11 indicate that these are'distinct groups genetically as well as geographically. A minor source of S. spontaneurn germplasm was the 2n = 80 form from Sulawesi, via the clone Toledo and the Tables indicate

8 B.T. ROACH 499 that the Sulawesi group is genetically distinct from other Indonesian S. spontaneum. The evolutionary origin of S. sinense/s. barberi remains uncertain, but its sampling via Chunnee and other clones of the Saretha and Pansahi groups in producing modern hybrids clearly added genetic diversity. This diversity may have had origins beyond Saccharum in Miscan thus saccharflorus (Grassl17) and Erianthus sect. Ripidium (Daniels12). The value of increasing genetic diversity from a wider sampling of S. spontaneum is supported by recent experience with use of S, spontaneum from Thailand. Heinz20 has reported high yielding progeny from this source while Roach and mull in^^^ reported high yielding progeny with a high level of waterlogging tolerance. The potential value of increasing genetic diversity beyond Saccharum is supported by reports of the drought tolerance of Erianthus sect. Ripidium hybrids (Walker50) and the ratooning ability and disease resistance of Miscanthus hybrids (Cheng). CONCLUSIONS The data on geographical origin and genetic diversity presented in this paper are very broad in nature but may be of some value in selecting wild germplasm for hybridisation. Within S. spontaneum, the north-east group of Indian clones appear generally different in some economic characters from those of other regions of India and thus merit sampling in breeding. The south-east Asian group and the Sulawesi and Kalimantan groups also deserve attention, particularly the latter because of its overlap for several characters into the S. robustum and S. officinarum ranges. Sampling of the other groups namely Japan/Taiwan, Western and Papua New Guinea is also merited, although our experience with hybrids of African and Papua New Guinea S, spontaneum has not been encouraging The situation in S. robustum is less clear, probably a reflection of the conclusion by Daniels and Roach13 that S. robustum is a name for some very diverse populations of plants derived from introgression of S. spontaneum with other genera in the Wallacea/New Guinea area. To sample the genetic variability of this "species", it would seem preferable to select representatives of each of the 5 groups defined by Price33 on a cytological basis and confirmed by Daniels" on a chemotaxonomic basis using leaf flavonoids. Erianthus arundinaceus, irrespective of its origin in Indonesia, exhibits relatively little variation. Berding7 has noted an apparent cline in flowering, with clones of Kalimantan flowering before those of Sulawesi which, in turn, flower before those from Irian Jaya. This may represent only minor genetic differences which do not merit multiple sampling in breeding in a search for heterosis. A large proportion of the genome of commercial hybrids is derived from S. " officinarum, although the exact proportion in unknown (R~ach~~). While the species has limited geographical distribution, it shows considerable variability. Where clones of the species are used rather than elite commercials as the nobilising parents, selection of the best available clones, or breeding of improved noble canes, as done by Walkers4g would seem appropriate. To make prompt use of selected germplasm from a range of sources, a deliberate

9 BREEDING program of bridging crosses would seem desirable, after the style of the (Badila x Tabongo) x (Badila x Mandalay) crosses referred to previously. This could be on the lines of a "population improvement" program outside the main breeding program, as suggested by Allison1. I ACKNOWLEDGEMENTS Staff of Macknade Experiment Station are acknowledged for assistance in collection of data, J. Daniels and R.T. Mullins for a critical reading of the manuscript and CSR Ltd for permission to publish this paper. ZL,, *( 8: it 1 t., ;" REFERENCES. 1. Allison, J.C.S. (1984). Use of new sugar cane germplasm. Sugar Cane. 4: Anon. (1982). Genetic resources of sugarcane. IBPGR working group on the genetic resources of sugarcane. IBPGR Secretariat, Rome. pp Arceneaux, G. (1965). Cultivated sugarcanes of the wdrld and their botanical derivation. Proc. ISSCT. - 12: Arunachalam, V. and Jawahar, Ram. (1967). Geographical diversity in relation to genetic divergence in cultivated sorghum. Indian J. Genet. P1. Breed. 27: Arunachalam, V., Bandyopadhyay, A., Nigam, S.N. and Gibbons, R.W. (1984). Heterosis in relation to genetic divergence and specific combining ability in groundnut. (Arachis hypogea L). Euphytica. 33: Berding, N. and Koike, H. (1980). Germplasm conservation of the Saccharum complex : A collection from the Indonesian archipelago. Hawaiian Planters' Record. 59(7): Berding, N. Personal communication. 8. Buddenhagen, I.W. (1977). Resistance and vulnerability of tropical crops in relation to their evolution and breeding. Annals New York Academy of Sciences. 287: Chen, W.H., Huang, Y.J., Shen, 1.S. and Shih, S.C. (1983). Utilisation of Miscanthus germplasm in sugarcane breeding in Taiwan. Proc. ISSCT. 18: Daniels, J. (1965). Improving sugarcane breeding methods to increase yields. Proc. ISSCT. 12~ Daniels, J., Smith, P., Paton, N.H. and Roach, B.T. (1977). The taxonomic status of Saccharum robustum. Brandes and Jesweit ex Grassl. Proc. ISSCT. 16: Daniels, J., Paton, N., Smith, P. and williams, Christine A. (1980). Further studies of leaf flavonoids as evolutionary indicators in Saccharum officinarum. Proc. ISSCT. 17: Daniels, J. and Roach, B.T. (1986). Taxonomy and evolution of sugarcane. In : Sugarcane improvement through breeding. ISSCT and Elsevier (in press). 14. Dunckelman, P.H. and Breaux, R.D. (1968). Evaluation of germplasm in the USDA sugarcane program at Houma, Louisiana. Proc. ISSCT. 13: Dunckelman, P.H. and Breaux, R.D. (1971). Breeding sugarcane varieties for Louisiana with new germplasm. Proc ISSCT. 14: Gill', S.S. and Tripathi, B.K. (1983). Nature of divergence among foreign varieties of sugarcane. Proc. ISSCT. 18: Grassl, C.O. (1964). Problems relating to the origin of wild and cultivated Saccharurn. Indian Journal Sugarcane Research and Development. 8: Hawkes, J.G. (1977). The importance of wild germplasm in plant breeding, Euphytica. 26: Heinz, D.J. (1965). Wild Saccharum species for breeding in Hawaii. Proc. ISSCT. 12:

10 B.T. ROACH Heinz, D. J. (1980). Thailand S. spontaneurn hybrid progeny as a new germplasm source in Hawaii. Proc. ISSCT. 17: Hogarth, D.M. (1971). Quantitative inheritance studies in sugarcane. I. Estimation of variance components. Aust. Jour. Agric. Res. 27: Hogarth, D.M., Wu, K.K. and Heinz, D.J. (1981). Estimating genetic variance in sugarcane using a factorial cross design. Crop Sci. 21: Irvine, J.E. (1965). Testing sugarcane varieties for cold tolerance in Louisiana. Proc. ISSCT. 12: Irvine, J.E. (1977). Identification of cold tolerance in Saccharum and related genera through refrigerated freeze screening. Proc. ISSCT. 16: Lo, C.C., Chen, W.H., Shen, I.S. and Shih, S.C. (1980). Reaction of Miscanthus germplasm to downy mildew and culmicolus smut of sugarcane. Proc. ISSCT. 17: Marimuthammal, S., Devayani, K.R. and Sankaranarayan, P. (1976). Studies in Saccharurn spontaneum L. Evaluation of bigh sugared forms. ISSCT Sugarcane Breeders' Newsletter. 37: Meade, G.P. and Chen, C.R. (1977). Cane Sugar Handbook, 10th Ed. John Wiley & Sons, New York. 28. Morishima, H. (1969). Phenetic similarity and phylogenetic relationships among strains of Oryzaperennis estimated by methods of numerical taxonomy. Evolution. 23: Mullins, R.T. and Roach, B.T. (1985). Genetic origins of ash in sugarcane juice. Proc. Aust. Soc. Sug. Cane Technol. 1985: Nagatomi, S. and Oshiro, Y. (1983). Classifications for sugarcane wild germplasm by methods of numerical taxonomy. Proc, ISSCT. 18: Nair, N.V. and Somarajan, K.G. (1983). A note on screening sugarcane germplasm for high early sucrose. Indian Sugar. Sept. 1983: Panje, R.R. and Babu, C.N. (1960). Studies in Saccharurn spontaneurn : Distribution and geographic association of chromosome numbers. Cytologia. 25(2): Price, Sam. (1965). Cytology of Saccharum robusturn and related sympatric species and natural hybrids. USDA Agricultural Research Service, Technical Bulletin, Roach, B.T. (1968). Quantitative effects of hybridisation in Saccharurn officinarurn x Saccharurn spontaneum crosses. Proc ISSCT. 13: Roach, B.T. (1971). Nobilisation of sugarcane. Proc. ISSCT. 14: Roach, B.T. (1977). Use of Saccharurn spontaneurn in sugarcane breeding. Proc. ISSCT. 16: Roach, B.T. (1984). Conservation and use of genetic resources of sugarcane. Sugar Cane. 2: Roach, B.T. and Mullins, R.T. (1985). Testing sugarcane for waterlogging tolerance. Proc. Aust. Soc. Cane Technol. 1985: Robbelen, G. (1979). Transfer of quantitative characters from wild and primitive forms. Proc. Conf. Broadening Genet. Base Crops, Wageningen. 1979: Singh, S.P., Srivastava, J.P., Singh, H.N. and Singh, N.P. (1977). Genetic divergence and nature of heterosis in Okra. Indian J. Agric. Sci. 47: Singh, H.N. and Singh, S.B. (1980). Genetic divergence in sugarcane. Proc. ISSCT. 17: , Srinivasan, K. and Batcha, M.B.G.R. (1962). Performance of clones of Saccharurn species and allied genera under conditions of waterlogging. Proc. ISSCT. 11: Stalker, H.T. (1980). Utilisation of wild species for crop improvement. Adv. Agron. 33: Venkatraman, T.S. (1936). The sweetness of the wild sugarcanes of India. Agriculture and Live Stock in India

11 502 BREEDING 45. Walker, D.I.T. (1971). Utilisation of noble and Saccharum spontaneum germplasm in the West Indies. Proc. ISSCT Walker, D.I.T., MacColl, D. and Rao, P.S. (1977). Aspects of the use of Saccharum spontaneum in the West Indies programme. Proc. ISSCT. 16: Walker, D.I.T. (1978). International co-operation in exploiting germplasm. Sugarcane Breeders' Newsletter. 41: Walker, D.I.T., MacColl, D. and Rao, P.S. (1980). Aspects of the use of Saccharum spontaneum in the West Indies programme. Proc. ISSCT. 17: Walker, D.I.T. (1985). Breeding vegetatively propagated plants : sugarcane. In : A.I. Campbell, A.J. Abbott and R.K. Atkin (Editors). Improvement of Vegetatively Propagated Plants. Academic Press, London, (in press). 50. Walker, D.I.T. Personal communication. 51. Williams, J.T. and Damania, A.B. (1981). Directory of germplasm collections. 5 - Cacao, coconut, pepper, sugarcane and tea. IBPGR Secretariat, Rome. pp