Chromosome Variation in Taro, Colocasia esculenta: Implications for origin in the Pacific. D. J. Coates1, D. E. Yen2 and P. M.

Transcription

1 _??_ 1988 by Cytologia, Tokyo C ytologia 53: , 1988 Chromosome Variation in Taro, Colocasia esculenta: Implications for origin in the Pacific Accepted August 22, 1987 D. J. Coates1, D. E. Yen2 and P. M. Gaffey2 As a major cultigen in subsistence agriculture of the Pacific Islands, taro or Colocasia esculenta (L.) Schott. has long been postulated as an example of prehistoric plant dispersal b y man out of SE Asia. This seemed to be confirmed by a study of chromosome numbers of clonal varieties collected from the two regions (Yen and Wheeler 1968). Because only forms with the somatic number of 28 were found from the tropical area of Polynesia, it was posited that the original "population" of ancestral material transferred into Oceania was of this chromo - some number, and that the 2n=42 clones found in New Caledonia and New Zelaland were subsequent introductions at historic time levels. The ultimate origin of the species was indicated as India, for from there somatic numbers had been recorded in the literature indicating two chromosome number series of 2n=14, 28 and 42, and 2n=36 and 48. Thus, on a founder principle manner of diminishing variation, a west-to-east line of diffusion by human agency was indicated in the chromosome numbers. This reopening of the subject was occasioned by the recent awareness of the present in vestigators of an early recorded Colocasia taro as a gathered food of the Aborigines of tropical Australia (Maiden 1889). This extension of taro distribution, despite a number of subsequent notices, e.g., Golson (1971), Cribb and Cribb (1975), Crawford (1982), has not entered the ethnobotanical literature of Oceania. The question of relationship to the cultivated forms of the Pacific Islands may be amplified in terms of a number of considerations: i) Wild forms of C. esculenta are also present in Papua New Guinea on the mainland and on some offshore islands. ii) Personal communications on the field work of Betty Meehan, Rhys Jones, Neville White and Neville Scarlett suggest that the Australian feral representatives of remote areas in the tropical north are of great antiquity. This contention is strongly supported by the fact that in some aboriginal groups the plant occupies a place in "dreaming", a fundamental l cultural phenomenon not normally associated materially with external introductions (R. Jones, pers. comm.) iii) The rising sea levels that separated the Sahul continent into the two land masses of New Guinea and Australia at the end of the Pleistocene some 11,000 years ago (Jennings 1971) had biological consequences one of which could have been the division of a formerly shared wild taro population. iv) Golson's structural evidence for agricultural drainage in the Papua New Guinea Highlands swamps beginning at some 9,000 years ago (Golson and Hughes 1980) points to earlier domestication-and to taro as a primary example. Since this indicates agriculture as early as any in the world, with therefore attendant difficulties of correlation with any putative diffusion, a hypothesis for the endemic, independent formation of agriculture has been proposed (Yen 1980). 1 Department of Population Biology, Research School of Biological Sciences, Australian National Uni versity. Present address: Wildlife Research Centre, P. O. Box 51, Wanneroo 6065, Western Australia. 2 Department of Prehistory, Research School of Pacific Studies, Australian National University, Aus tralia.

2 552 D. J. Coates, D. E. Yen and P. M. Gaffey Cytologia 53 The investigation of the "new material" from Australia and its relationship to the tax onomically undifferentiated Pacific representatives of the species then, returns to the cyto logical approach. Previous cytological investigations on taro, particularly on materials from India, indicate some confusion concerning the basic chromosome number within the species. Darlington and Wylie (1955) characterised Indian variants as possessing two basic numbers, x=12 and 14, the former determination based on the observations of Delay (1951) and Rao (1947). However none of the more recent studies on Indian material (Sharma and Das 1954, Mookerjea 1955, Pfitzer 1957, Vijaya Bai et al. 1971, Kuruvilla and Singh 1981), including the survey of Ramachandran (1978) of 17 cultivars from four regions of India, have confirmed x=12 as a base number. Further chromosome number surveys on taro from the Pacific region, China and Japan (Yen and Wheeler 1968), 103 different cultivars from Japan (Fukushima et al. 1962) revealed only 2n=28 or 42 variants. Kawahara (1978) records the chromosome numbers of 13 cultivars from Nepal and two from India as 2n=42, while both 28- and 42-somatic chromosome numbers occur in five wild Nepalese forms. It therefore seems likely that the variants observed with a base number of 12 were either misidentified as taro, or that the chromosome counts were inaccurate. Although the cytological data to date strongly suggests that x=14 is the base number for Colocasia, Mookerjea (1955) has reported an Indian form with somatic complement of 14. The contention that this form may have been a rare haploid variant (Yen and Wheeler 1968) has support from the meiotic studies of Krishnan et al. (1970) and Vijaya Bai et al. (1971) on 2n=28 and 42 forms from India, although more recently, Krishnan and Magoon (1977), on evidence from a karyological study, advocate a base chromosome number of x=7 (see also Marchant 1971). Since the survey of Pacific taro by Yen and Wheeler (op. cit.), there has been one further record of 42 chromosome forms by Jackson et al. (1977) among the 2n=28 cultivars of the Solomon Islands. This extends the range of the triploid (or hexaploid) forms of Asia, the Philippines, Timor, New Caledonia and New Zealand. These cytological issues relating to Colocasia are of importance for the present study aimed at the further investigation of chromosomal variation of the species C. esculenta in the Pacific, particularly in view of the hitherto unknown forms in Australia. More detailed chromosome studies on the material were planned utilising karyotype data that might indicate variability both within and between expected 28- and 42- chromosome forms, and possible differentiation of wild and cultivated types. One recent example of this approach is the account of Kuruvilla and Singh (1981), in which one wild taro from India (2n=28) exhibited a more asymmetrical karyotype than 14 cultivars of which three were 28- chromosome forms and the remainder 42. Materials and methods The material for cytological analysis (Table 1) was collected in two ways: from field sites in Australia, New Zealand, Papua New Guinea, the Philippines, Thailand, Japan, Nepal, Huahine in the Society Islands and Easter Island as live plants for propagation in glasshouses at the Commonwealth Plant Quarantine Research Station in Canberra; from the Lyon Arboretum taro collection, University of Hawaii, as directly pretreated and fixed root tip material. The pre-treatment of extracted root tips in both cases was immersion in 0.2% aqueous col chicine for four hours followed by fixation in 3:1 absolute alcohol; glacial acetic acid at room temperature, and storage at 4 Ž. The material was prepared for chromosome studies by the Feulgen technique; hydrolysis for 4 to 5 minutes in 1N HCI at 60 Ž, and squashing in aceto orcein. Suitable mitotic metaphases were photographed at ~ 1,000 under oil immersion, and the resultant prints enlarged for karyotype assembly.

3 1988 Chromosome Variation in Taro, Colocasia esculenta 553 Results The chromosome numbers of the 107 Pacific (including the Philippines, Japan and Thailand) island clonal varieties do not deviate from the pattern set in the earlier work of Yen and Wheeler (1968), demonstrated in Table 1. Most are 2n=28, while the 2n=42 forms are represented in the taro assemblages from Timor, the Philippines, Australia, New Zealand, Japan, Nepal and India. It may be recalled from the assessments of Mookerjea (1955) and more recent work on Indian material that most cultivars from that country are of the 42-form. The material collected for this study from Papua New Guinea, including separate wild clones, Philippines, Palau, Society Islands is not duplicated in the Yen and Wheeler (1968) study, and thus provides reinforcement to the contention of primacy of the 28-form in the Pacific islands. Table 1. Summary of collection area of Colocasia esculenta, the chromosome numbers and the number of samples karyotyped in this study b From Lyon Arboretum, remnant of collection of Yen and Wheeler (1968). b Collected from Lyon Arboretum, University of Hawaii. c From Botanic Garden, Brisbane, Australia, no provenance. d New Zealand collections, P. Matthews (1985). e From five separate wild populations, Northern Territory, Queensland. f From nine separate wild populations, Highlands, Wahgi Valley; Baiyer River; Tun Valley; Rulna, Lowlands, Goldie River, Papua, Island, Moa Is., Torres Strait. With regard to the Australian clones, there are two categorised as cultivated and with the 42 chromosome number. As unprovenanced plants grown in a botanical garden, they could well be introduced from New Zealand or New Caledonia or even India. Of significance, however, are the six of seven remaining Australian clones designated as "wild", collected from localities in Queensland and the Northern Territory (some sites being quite remote, e.g., in the Mitchell Range), all with 2n=28 complements. From the total material of 112 clonal samples, 81 were karyotyped (see Table 1). This analysis revealed that the 28-form consists of two cytotypes (Figs. la and b) while three cytotypes were observed in the 42-form (Figs. 2a, b and c). Cytotype I, illustrated in Fig. la, represents the majority of plants karyotyped including both wild and cultivated clones. The somatic chromosome complement as arranged conforms with the hypothesis that the species

4 554 D. J. Coates, D. E. Yen and P. M. Gaffey Cytologia 53 has a base number of 14, characterised by 14 submetacentric pairs of chromosomes. For con venience then, we refer to the 28 somatic chromosome form as diploid and the 42 as triploid. The karyotypes of four diploid plants, two from Queensland, Australia (one collected from rain forest near Mossman, Cape York, the other west of Brisbane) and two cultivars from the west highlands of Papua New Guinea reveal a chromosome complement identical to cytotype I except that chromosome pair 7 is heterozygous for a pericentric rearrangement resulting in metacentric and acrocentric forms of chromosome 7 within the same plant (Fig. 1b). Although this deviancy would generally be considered to be a polymorphic variant of cytotype Fig. 1. Karyotypes of the two diploid (2n=28) cytotypes: (a) cytotype I, (b) cytotype I-1. Marker chromosomes 3, 7 and 9 are underlined. I, for discriminatory reference it is here designated as cytotype I-1. It should be noted that none of the diploid plants karyotyped were homozygous for the derived acrocentric form of chromosome 7. The analysis of the triploid 42 clones indicate autotriploidy, an interpretation supported by the results of meiotic studies of Krishnan et al. (1970) and Sharma and Das (1954). However, the karyotypes can be differentiated into three cytotypes: cytotype I-3 characterised one cul tivar from Luzon, Philippines and one from Timor. This type illustrated in Fig. 2a, is so designated since it appears to be an autotriploid derived from cytotype I; the haploid karyotype is identical to that found in the diploid form shown in Fig. la. Cytotype I-13, found only in four clones of the New Zealand material (Fig. 2b) can be derived directly from cytotype I-1

5 1988 Chromosome Variation in Taro, Colocasia esculenta 555 Fig. 2. Karyotypes of the three triploids (3n=42): (a) cytotype I-3, (b) cytotype I-13, (c) cytotype II-3. Marker chromosomes 3, 7 and 9 are underlined. Note in cytotype II-3; all three marker chromosomes are acrocentric.

6 556 D. J. Coates, D. E. Yen and P. M. Gaffey Cytologia 53 (Fig. 1b), although it is unclear how it arose given the geographic distribution of I-1, on present knowledge confined to Australia and Papua New Guinea. The karyotype is the same as that found in the autotriploid I-3 (Fig. 2a) except for chromosome 7 where there are two copies of the derived acrocentric chromosome found in I-1 and one copy of the standard metacentric chromo some. Cytotype II-3, represented by the triploids, from Australia, New Zealand, India, Nepal and Japan, cannot be derived directly from either of the diploid cytotypes found in this study (Fig. 2c, cf. Figs. 1a, 1b). Karyotypically, this cytotype is quite different from all of those designated as I. Obvious differences involve chromosomes 3 and 9 which are acrocentric rather than metacentric, while chromosome 7 is now homozygous acrocentric rather than het erozygous as in the diploid I-1 (Fig. 1b) and the triploid I-13 (Fig. 2b). Discussion The present cytological investigation on taro confirms previous chromosome studies in dicating that the species is characterised by two somatic chromosome numbers 28 and 42, con sidered to be diploid and triploid respectively. Within the Pacific region, the findings of Yen and Wheeler (1968) regarding the distribution of these forms, particularly the representatives in tropical Polynesia and Papua New Guinea as exclusively 2n=28, have not been contradicted. On numerical data alone, however, the recent record of the triploid from on the Solomon Islands (Jackson et al. 1977) gives impetus to the question of provenience of the New Zealand forms of 2n=42. Explanations of direct transmission in prehistory from Melanesia to New Zealand are generally unacceptable on cultural grounds, so that the alternatives remain of introduction in historic times as Yen and Wheeler (1968) proffered, or that 42 chromosome forms arose independently in the Pacific from 28-forms. That the representatives of the au thenticated wild populations of Australian Colocasia esculenta are all of the somatic comple ment of 28 may be a finding of importance in considering the phylogeny of the species in the Pacific. The karyotype studies reported herein show that the 28 chromosome forms found through out the taro investigated from the Pacific islands and Australasia and including samples from Micronesia, the Philippines and Timor, have the same diploid chromosome complementcytotype I. It seems likely that this major Oceanic cytotype was present in Australia and Papua New Guinea prior to cultivation by man, since it survives there still as both wild and domesticated forms of the species. The 28 chromosome cytotype designated as I-1, hetero zygous for a pericentric rearrangement in chromosome 7, and represented in two feral Aus tralian clones and two New Guinea cultivars, are seen as most likely to have been descendant from cytotype I. The rearrangement, whether by inversion or centric transposition, can be assumed to be a rare event, with heterozygosity (no homozygous acrocentric complements were found) persisting through natural, and subsequently on domestication, by horticultural vegeta tive propagation. Further, the limited distribution of this variant within the range of Colocasia in the Pacific is suggestive of an Australian-New Guinea origin. Of the three triploid forms (2n=42), the karyotypic analyses suggest that the clones homo zygous metacentric for chromosomes 3, 7 and 9 (cytotype I-3) could have been directly derived from cytotype I by autotriploidy. Its distribution, restricted to Timor and the Philippines, suggests that this variant is of SE Asian origin. Cytotype I-13 may be most economically viewed as an autotriploid descended from cytotype I-1 by duplication of the acrocentric chro mosome in the heterozygous chromosome pair 7. Its distribution, being so far found only in New Zealand, poses a problem of place of origin, unless it can be held that this chromosome form is a local development of obligatory recency, New Zealand being colonized primarily at c. A. D. 800 by the Maori who presumably brought taro. The interpretative difficulty, how-

7 1988 Chromosome Variation in Taro, Colocasia esculenta 557 ever, goes beyond this, for the accepted immediate provenance of the New Zealand Maori is Eastern Polynesia (see Davidson 1984 as an example of many possible references.) The putative parent for cytotype I-13 is found in Australia and New Guinea; proposals for direct prehistoric contacts of New Zealand with either would hardly have archaeologists' support from their evidence of unrelated prehistoric material cultur es.th e remaining triploid form designated as II-3 is karyotypically quit e distinct, with all three indicator chromosome groups 3, 7 and 9 being homozygous acrocentric. Clones from N ew Zealand, India, Nepal, Japan and two questionable cultivars from the Brisbane Botanic Gardens exhibited this karyotype while there is no obvious diploid ancestor in ou r material. Fig. 3. Proposed phylogenetic relationship between the five cytotypes based on karyotype and ploidy level; whether the marker chromosomes 3, 7 and 9 are metacentric (m) or acrocentric (A) and the number of copies of each type are indicated for the different cytotypes. We are left with the hypothesis and these triploid forms are relatively recent introductions from Asia-an explanation offered for the New Zealand and New Caledonian presence of 42- chromosome forms by Yen and Wheeler (1968). If the pattern of autopolyploid ancestry is applied to cytotype II-3, the likely pattern would be a diploid parental cytotype homozygous acrocentric for chromosomes 3, 7 and 9 that could be designated as hypothetical cytotype II. There is indeed some indirect support for the presence of the diploid cytotype II in India from the studies of Kuruvilla and Singh (1981). Although requiring confirmation, the chromosome spread of a 2n=28 form presented in that paper (Fig. 2.1) indicated a number of acrocentric

8 558 D. J. Coates, D. E. Yen and P. M. Gaffey Cytologia 53 chromosomes. It was not clear whether three acrocentric pairs of chromosomes were present, as postulated for cytotype II (Fig. 3), although it is evident that the karyotype would undoubtedly be different from that of the widespread diploid cytotype I. Whatever the origin of cytotype II-3, it is clear that there are two distinct lineages which characterise taro presently found in the region of Australasia and the Pacific. The proposed scheme for this phylogeny is presented in Fig. 3. The application of karyotypic analysis to the Pacific taro material has been able to produce a hypothetical divergency in the species that could not be produced by the earlier numerical approach to cytological study. The presence of the triploid cytotype II in Australasia and its absence in the rest of Oceania suggest an ex ogenous source that we have identified as the Indian subcontinent from restricted comparative material (see Table 1). It is likely, however, that such cytotypes (including perhaps diploid II forms?) may be distributed on the SE Asian mainland or Indonesian islands. The sources of the material investigated nearest to these regions, however, are from Timor and the Philippines, and although two triploids are found, they are both of the I lineage. The I lineage (Fig. 3) has its Pacific foundation in the widespread diploid chromosome form of homozygous metacentrics, out of which two forms of triploids would seem to have arisen, one by direct autopolyploidy, the other through an intermediate diploid stage following pericentric rearrangement in one chromosome pair. That the putative progenitor cytotype I (with its derivative cytotype I-1) are found in the wild populations of Australia and New Guinea can be looked on as evidence for considerable antiquity for the form in the region. Indeed, rather than as part of human dispersal of agriculture from Asia at c. 5,000 years ago (Bellwood 1978), the initial appearance in the western Pacific of Colocasia taro, in its cytotype I form, was likely within the eastern extension of the Indo-Malaysian flora with the collision of the Australia- New Guinea or Sahul continent with the Asian plate. Palaeogeographic research, e.g., Audley- Charles et al. (1981), estimates the beginning of this process of floral exchange at between the middle Miocene and the Pleiocene. Thus the temporal, environmental and chromosomal cir cumstances exist for the hypothesis of a separate domestication of taro in the western Pacific -supported by archaeological evidence for the species in agriculture as early as any reported in the world. Conclusion In this study of taro cytology emphasizing the karyotypic approach, the Indian-SE Asian mainland origin of at least one form of the species remains unquestioned. However, the results indicate that a natural distribution concurrent with late Tertiary geological events may more satisfactorily account for the spatial differentials in chromosome number and cytotype dis tribution than the accepted theory of dispersal by human diffusion at some 5-6,000 years ago. The acceptance of such a view requires a further extension-that the western Pacific taro was the subject of a separate domestication process parallel to that on the Asian continent. It is nota ble that the majority of the cultivated and wild forms of taro in the Pacific are of the 2n=28 chromosome complement, homozygous metacentric for three diagnostic chromosomes; as cultivars, this form, designated in this work as cytotype I, was distributed eastwards by man to the far reaches of Polynesia. In the western region moreover, the karyotypes of the taro material suggests that diploid cytotype I was the progenitor in the transformation to triploid forms through direct autopolyploidy in one instance, and through an intermediate step of pericentric rearrangement in the other. The finding of the latter triploid type (I-13) so far only in New Zealand poses a question of its geographic provenance. The karyotypic analysis of the New Caledonian and Solomon Islands triploids, not carried out in this investigation, is an obvious step, although any equivalence of that material with the New Zealand I-13, as has

9 1988 Chromosome Variation in Taro, Colocasia esculenta 559 already been indicated, would encounter explanatory difficulties of a prehistoric cultural nature. However, for the wider implications of the present work which indicate a divergence of evolutionary pathways of the Pacific lineage I and the II lineage (represented here only by the triploid form II-3) determination of the karyotypes of Indian and SE Asian taro over a wide.geographic range may provide the ultimate resolution. Summary Further cytological study of Colocasia taro in the Pacific utilizing karyotypic data has produced a hypothesis for two separate lineages of the plant within contemporary populations. This provides support for domestication in the western Pacific independent of diffusion of Asian cultigen forms, but it is indicated that confirmation awaits more comprehensive karyotyping of Indian and Southeast Asian material. Acknowledgements We thank the Department of Primary Industry, Honiara, for additional material from the Solomon Islands, and the Kyoto Plant Germ Plasm Institute at Kyoto University for varieties from Japan and Nepal. References Audley-Charles, M. G., Hurley, A. M. and Smith, A. G Continental movements in the mesozoic and Cenooic. In: T. C. Whitmore (ed.), Wallace's Line and Plate Tectonics, pp Clarendon Press, Oxford. Bellwood, P. S Man's Conquest of the Pacific. The Prehistory of Southeast Asia and Oceania. Oxford University Press, New York. Crawford, I. M Traditional Aboriginal Plant Resources in the Kalumburu Area: Aspects in Ethnoeconomics. Western Australian Museum, Perth. Cribb, A. B. and Cribb, J. W Wild Food in Australia. Collins, Sydney. Darlington, C. D. and Wylie, A. P Chromosome Atlas of Flowering Plants. George Allen and Unwin Ltd., London. Davidson, J The Prehistory of New Zealand. Longman Paul, Auckland. Delay, C Nombres chromosomiques chez les phaneorgames. Rev. Cytol. Biol. Vegetales 12: Fukushima, E., Iwasa, S., Tokumasu, S. and Iwamasa, M Chromosome numbers of the taro varieties cultivated in Japan. Chromosome Information Service 3: Golson, J Australian aboriginal food plants; some ecological and culture-historical implications. In: D. J. Mulvaney and J. Golson (eds.), Aboriginal Man and Environment in Australia, pp Australian National University Press, Canberra. - and Hugheș P. J The appearance of plant and animal domestication in New Guinea. Journal de la Societe des Oceanistes 36(69): , Paris. Jackson, G. V. H., E A. Ball and Arditti, J Tissue culture of taro, Colocasia esculenta (L.) Schott. J. Hort. Sci. 52: Jennings, J. N Sea level changes and land links. In: D. J. Mulvaney and J. Golson (eds.), Aboriginal Man and Environment in Australia, pp Australian National University Press, Canberra. Kawahara, T Chromosome number of taros in Nepal and India. Chromosome Information Service 24: 4-5. Krishnan, R. and Magoon, M. L Edible aroids-new insights into phylogeny. In: C. L A. Leakey (ed.), Proc. 3rd Int. Sump Trop. Root Crops, pp Int. Inst. Tropical Agric., Ibadan, Nigeria. -, - and Vijaya Bai, K Desynapsis in Colocasiantiquorum Schott. Genetica 41: Kuruvilla, K. M. and Singh A., Karyotypic and electrophoretic studies on taro and its origin. Euphytica 30: Maiden, J. H The Useful Native Plants of Australia (Including Tasmania). Trubner and Co., London. Marchant, C. J Chromosome variation in Araceae II. Richardieae to Colocasieae. Kew. Bull. 25(1):

10 560 D. J. Coates, D. E. Yen and P. M. Gaffey Cytologia Matthews, P Nga taro o Aotearoa. Journal of the Polynesian Society 94(3): Mookerjea, A Cytology of different species of aroids with a view to trace the basis of their evolution. Caryologia 7: Pfitzer, P Chromosomenzahlen von Araceen. Chromosoma 8: Ramachandran, K Cytological studies on South Indian Araceae. Cytologia 43: Rao, N. S A note on the chromosome number of Colocasia antiquorum Schott. Curr. Sci. 16: 229. Sharma, A. K. and Das, N. K Study of karyotypes and their alteration in Aroids. Agronomia Lusitania 16: Vijaya Bai, K., Magoon, M. L. and Krishnan, R Meiosis and pollen mitosis in diploid and triploid Colocasia antiquorum Schott. Genetica 42: Yen, D. E The southeast Asian foundations of Oceanic agriculture: a reassessment. Journal de la Society des Oceanistes 36(66-67): , Paris. - and Wheeler, J. M Introduction of taro into the Pacific: the indications of chromosome numbers. Ethnology 7: