Robertsonian fusion and centric fission in karyotype evolution of higher plants.
Eukaryotes show wide variation in the number and morphology of chromosomes in their complements. These are known and generally described from observations of mitotic chromosomes whose numerical and morphological aspects comprise the karyotype. For the great majority of organisms this represents all that is known of their cytology, the karyotype becoming increasingly prominent as the favorable target for painting. That fact, coupled with the frequent sacrifice of the chromosome to the DNA amount of the nucleus, leads inevitably to the neglect of meiosis and loss of the information it provides on genome relationships and on mechanisms and pathways of chromosome evolution. In this review of two closely interrelated types of structural change involved in the evolution of karyotypes, the relative importance of mitotic and meiotic observations will become apparent.
Robertsonian fusion and centric fission are unique in being readily identified in comparative karyotype studies since both result in concomitant changes in chromosome morphology and chromosome number. Essentially, their involvement becomes apparent when a metacentric in one complement is replaced by two acro- or telocentric chromosomes in another, though that observation does not in itself distinguish between the two processes. Meiotic analysis can be critical in that analysis, as shown particularly in the Tradescantieae (vide infra).
The importance of these kinds of change in karyotype evolution is shown by their high incidence in animals, occurring in molluscs, insects, reptiles, and mammals either as species-specific differences or less commonly intraspecific polymorphisms (cf. White, 1973; King, 1993). In higher plants, in contrast, such changes have been detected in relatively few genera and consequently have not been thought to have played such a major role.
The mechanisms underlying both fusion and fission have been described many times, principally by White (1945, 1973) with modifications by John and Freeman (1975). Briefly, fusion is the consequence of breakage proximal to or within the centromere of two usually nonhomologous acro- or telocentric chromosomes followed by reunion uniting the longer and shorter segments, respectively. The large metacentric possesses all of the significant genetic information of the participating chromosomes. The small metacentric is assumed to be genetically insignificant and, because of its drastically reduced size, meiotically incompetent and liable to be rapidly eliminated from the breeding population. [Cases where it is not dispensable and survives, as, for example, spontaneous mutants in the Aloineae (Brandham, 1983), do not come within the strict definition of fusion of the Robertsonian type.] Chromosome number is then reduced and karyotype symmetry increased without a change in the number of major chromosome arms (= nombre fondamantal or NF).
Centric fission is the consequence of a break within the centromere of a single metacentric producing initially two structural telocentrics whose raw ends have the capacity to fuse after replication. This process produces an isochromosome, but where the mutation succeeds, telomere sequences are added to the ends. This in turn gives rise to two stable chromosomes of telocentric appearance. Amplification of these sequences has the ability to eventually create a detectable and heterochromatic short-arm. Fission increases chromosome number and karyotype symmetry; NF remains constant.
Figure 1 shows a simplified interrelationship of fusion and fission when fusion occurs through the centric region only. A preference for this site is indicated in Brandham's (1976) observations of spontaneous chromosome mutations and accords with Redi's (Redi et al., 1990) proposals for a likely mechanism of breakage and recombination in that region which seems to have sequences homologous with those of other centromeres across the complement.
With relation to both mechanisms, the terms "acrocentric" and "telocentric" were generally interchangeable when first proposed, expressing contrasting opinions on the presence of a second but generally invisible short-arm in the first and of a terminal centromere in the second. The techniques for identifying telomere sequences have shown the ephemeral existence of terminal centromeres after breakage and demonstrate that stable chromosomes of telocentric morphology have a telomere at both ends (Werner et al., 1992; Cox et al., 1993). The term "telocentric," however, is useful in a purely morphological sense, as in the proposed system of chromosome terminology of Levan et al. (1964). That system will be used in this review when attempting to transpose the karyotype descriptions of authors for comparative purposes. In doing that, although the abbreviations "M" and "ms" will differentiate chromosomes with equal or somewhat unequal arms, respectively, they will sometimes be grouped as metacentrics; in a similar way, "st" and "t" chromosome types with detectable short-arms of varying size may be grouped together as acrocentrics, leaving "T" exclusively for chromosomes with no visible second ann.
When discussing the significance of differences resulting from single or successive fusions or fissions it is common to ascribe their importance solely to postmating barriers to gene exchange and the progressive isolation of populations and incipient species that result (cf. King, 1993). In addition, however, it needs to be stressed that both processes must have profound influences on genetic recombination and segregation by changing the number and nature of linkage groups (interchromosome modification), while the known marked differences in frequency and distribution of chiasmata in metacentrics as compared with acro- or telocentrics (Mattsson, 1971; Bideau, 1990, 1993) should result in different patterns of intrachromosome recombination. An awareness of all these consequences of basically simple structural changes makes it possible to understand their frequent utilization in animals, and also makes it difficult to believe that they have been involved only rarely in higher plants' evolution. However, as we shall see, they are detectable only when the diagnostic features of their presence - i.e., the metacentric-acrocentric relationship and constant NF - remain. In plants, further structural changes, including those resulting from recombination in species of hybrid origin and to some extent polyploidy, can combine to remove or obscure both. Furthermore, fusion and fission are detectable only when chromosomes are large enough to display the features that identify them. The majority of taxa have chromosomes that fall within the small or minute range of size, where both the analysis of chromosome form and the calculation of NF may be impossible, or at best unreliable.
Finally, before describing investigations of adaptive fusion and fission, it is useful to recall that these and other chromosomal mutations are found as sporadic spontaneous mutations in natural populations in frequencies as high as 2%, providing an ample supply of variants for exposure to natural selection (Brandham, 1976; Parker et al., 1988). Spontaneous fusions have been seen in many taxa, including the Aloineae (Brandham, 1983) and Lycoris (Kurita, 1987c); and fissions, in such taxa as Hypochaeris (Parker, 1987), Nigella (Strid, 1968), Tradescantia (Jones & Kenton, 1984), and Vicia (Schubert & Reiger, 1990). In the latter the fission products are invariably of telocentric appearance.
III. The Cases
The New World genera of the Commelinaceae which form this subfamily, and particularly those where whole-arm rearrangements were identified, have been subjected to close cytological scrutiny. Subsequent to the publication of several reports on these studies, Hunt (1980) reunited a number of chromosomally distinct genera including Zebrina and Cymbispatha with Tradescantia, allocating them to a range of series and sections.
1. Tradescantia sect. Zebrina
Two species investigated as Zebrina - i.e., Z. flocculosa 2n = 22 (6 M + 6 st + 2 t + 8 T) and Z. pendula 2n = 24 (4 M + 6 st + 2 t + 12 T) - differ by a single homozygous fusion or fission (Mattsson, 1971). Mattsson assumed both species to be tetraploid evolving from a diploid base with x = 6M with subsequent considerable loss of major arms. That unlikely event can be eliminated by assuming that the common NF = 28 indicates a base of x = 7 and that the present complements of the species derive from a series of fusions starting with an ancestral diploid constitution of 2n = 14 st, t, or T chromosomes or a mixture of all three. The present complements of the two species represent successive stages in that progression though a reversion from 2n = 22 to 2n = 24 by fission cannot be excluded. Support for the general hypothesis comes from the cytology of an ally now called Tradescantia soconuscana and a diploid closely related to it (Jones, 1990).
Tradescantia soconuscana, 2n = 26 (6 M + 16 st + 4 T), although having a diploid number, rare quadrivalent formation within each chromosome class coupled with NF = 32 (4 x 8), suggests a fundamentally tetraploid constitution - the implication being that the M chromosomes are fusion derivatives from an earlier constitution of 2n = 32 st or T chromosomes. That possibility is given credence by the discovery of a close relative with 2n = 16 (8 st + 8 T). Assuming an allopolyploid origin for T. soconuscana, the hybridization of two diploids - one homozygous for a single fusion (2 M + 8 st + 4 T) and the other for two (4 M + 8 st) - followed by chromosome-doubling would result in the present complement. With that interpretation, fusion would have been confined to chromosomes of the T type.
2. Tradescantia sect. Cymbispatha
Members of this section are thought to be distantly allied with those of sect. Zebrina. Within Tradescantia commelinoides four cytotypes - 2n = 14 M, 2n = 22 (20 M + 2 t), 2n = 28 M, and 2n = 30 (26 M + 4 t) - have NFs = 28, 42, and 56 (Jones, 1977, 1978). The apparent diploid with 2n = 14 M forms quadrivalents in frequencies expected in an autotetraploid; the 2n = 22 is meiotically autohexaploid and the 2n = 28 and 30 types, auto-octoploid. These observations, coupled with the NF values, point to the conclusion that they are, genetically speaking, closely related tetraploid, hexaploid, and octoploid races with chromosome numbers changed by successive fusions. Another species outside the T. commelinoides complex, T. standleyi is 2n = 16 (12 M + 4 t) and forms bivalents only, but with highly proximal chiasmata inhibiting any potential multivalent formation. It, too, is considered genetically tetraploid on the basis of its NF = 28 and chromosomes occurring as groups of four homomorphs. Another species, T. poellii 2n = 36 (6 M + 30 t), is also bivalent-forming, and, with NF = 42, is "hexaploid" possibly of hybrid origin.
The origin of these "polyploids" must lie in diploids with constitutions like those found in two species: T. gracillima, 2n = 14 t, and T. plusiantha, 2n = 12 (2 M + 10 t). Although not considered to be ancestral species of the T. commelinoides complex, their differentiation by a single homozygous fusion testifies to the operation of fusion at that level in the genus.
Successive fusions proceeding to the maximum possible extent would produce the constitution 2n = 8 (6 M + 2 t). An autotetraploid derivative could move to complete metacentricity, but only by fusions between homologues producing when homozygous a pair of pseudoisochromosomes. Support for that inevitable conclusion was provided in one 2 n = 14 M Tradescantia commelinoides individual with abnormally low chiasma frequency which frequently formed one and rarely two ring univalents typical of internal pairing of chromosomes of that type. It confirmed the assumption of origin of the 14 M constitution while at the same time implying that isochromosomes can be stable members of normal complements when genetic regulation excludes internal pairing. The identification of isochromosomes can then be a valuable indication of possible tetraploid ancestry. That subject has been discussed in detail elsewhere (Jones, 1978).
Although the cytology of the T. commelinoides group has been dealt with in a number of publications, the full details upon which the above conclusions have been made appear in only one (Jones et al., 1981).
3. The Gibasis karwinskyana Alliance
Two cytoraces of Gibasis scheideana, 2n = 10 (4 M + 6 st) and 2n = 16 (12 M + 4 st), behave meiotically as diploid and autotetraploid, respectively. FI triploids showed close homology between parental sets while pairing between M and sts was of the type expected had the latter been involved in fusion. It is assumed that this produced the change from x = 5 to x = 4 at the diploid level prior to chromosome-doubling. B chromosomes found in several tetraploids are of similar morphology to the short-arm lost in fusion (Jones, 1974).
Within the group to which Gibasis scheideana belongs, both G. karwinskyana and G. consobrina have diploid 2n = 10 and tetraploid 2n = 20 forms sharing a morphologically similar set of x = 5 (2 M + 3 st), but G. matudae and G. pulchella, both 2n = 10, are distinctive in the genus with their complements composed exclusively of large M and sm chromosomes (Jones et al., 1975) and DNA amounts per genome equal to 17.78 and 21.05, respectively - double the amounts known for diploids in the other species, most lying within the range 8.05-11.41 (Jones & Kenton, 1984). In a detailed analysis of population samples of G. pulchella, the C-banding characteristics of all 10 chromosomes were identified. Meiosis was examined in bivalent-forming chromosomes and those with interchange-multivalents, including self-fertile populations that were true-breeding for complete rings of all ten chromosomes and therefore, in all respects, complex interchange-heterozygotes (Kenton et al., 1987, 1988). In a number of individuals with reduced chiasma frequency, both ring-univalents and fold-back bivalents (i.e., those where one arm of a metacentric formed chiasmata both with its second arm and with its homologue) were found, indicating the presence of duplicate terminal segments within single chromosomes, which in turn might suggest the presence of a pair of isochromosomes.
In combination, chromosome characteristics and DNA amounts for the 2n = 10 M species could be interpreted as evidence for an origin by fusion and chromosome-doubling from an ancestral base of 2n = 10 t - though that was not the preferred explanation of Kenton and her colleagues, who favored a purely diploid constitution for both.
4. The Gibasis linearis Alliance
In this alliance, species exist with x = 5 (3 M + 2 sm) and x = 6 (2 M + 3 sm+ 1 st) at diploid and tetraploid levels. In one 2n = 22 tetraploid, combining both basic numbers a single st-MM-st quadrivalent shows the homology between the constituent chromosome types (Jones & Bhattarai, 1981). Detailed investigation of artificial hybrids at diploid, triploid, and tetraploid levels further confirmed the homology between st and M arms (Kenton, 1981a, 1982, 1984). Overall it was concluded that variation in karyotype pattern in this group of species was probably initiated by a single fusion between chromosomes which are now sm and st but which became modified by inversion and translocation subsequent to fusion.
The above studies underline the value and importance of population sampling and of meiotic analysis in species and synthetic hybrids for the detection of fusion. In some cases, measurements of DNA amount and analysis of C-band patterns have made useful contributions. In total they show the interaction of fusion and chromosome-doubling in producing chromosome numbers which, without the benefit of meiotic observations, conceal fundamental genomic constitution. An awareness of isochromosomes and their implications has also been shown to be informative in detecting hidden polyploidy. The frequency of fusion across a range of species and genera identifies it as a type of chromosome ortho-selection, making likely its involvement in the evolution of complements in other taxa in the subfamily.
This bulbous genus of the Amaryllidaceae was one of the earliest recorded cases of whole-arm rearrangement. The first studies of Inariyama, in 1931, were followed by many others including an extensive series of papers by Kurita (1986, 1987a, 1987b, 1987c, 1988a, 1988b) which provides a complete bibliography on the subject.
A small group of diploid species with a centre of distribution in subtropical southern China have karyotypes considered to be ancestral for the genus as a whole (Kurita, 1988b). Lycoris chinensis, L. longituba, and L. straminea are 2n = 16 (6 M + 10 t). Lycoris traubii is variable, with 2n = 14 (8 M + 6 t), 2n = 13 (9 M + 4 T), and 2n = 12 (10 M + 2 t), the 2n = 13 form being sterile. Japanese populations of L. traubii were entirely 2n = 14 or a mixture of 2n = 12 and 13 types in similar proportions. Each of the species and their cytological variants have NF = 22 and M chromosomes twice the length of the t chromosomes, and are the largest found in the genus. Inariyama (1951) observed a t-M-t trivalent in the 2n = 13 L. traubii. These data indicate that fusion or fission is responsible for the observed karyotypic variation.
Another group of hardier species of the northern temperate zone with assumed derivative karyotypes - Lycoris albiflora, L. radiata, L. sprengerii, L. sanguine, etc. - are 2n = 22 st, 2n = 33 st, and rarely 2n = 44 st, sharing with the previous group NF values that are multiples of 11. The st chromosomes gently grade in the length of their long-arms; the substantial short-arms are of similar size. In general they are shorter in overall length than the t chromosomes of the previous group. DNA content of the M + t genomes is larger than that of the 2n = 22 st species.
A third group now widespread in Korea and Japan through human intervention, are sterile hybrids combining the karyotypes of the first two groups in various diploid and polyploid combinations. Some of these provided the opportunity to assess structural homology of the parental genomes by meiotic analysis (Inariyama, 1932; Koyama, 1962, 1978). Lycoris albiflora 2n = 17 (5 M + 11 st + 1 t), a probable hybrid of L. traubii 2n = 12 (10 M + 2 t) and L. radiata 2n = 22 st, had up to five st-M-st trivalents in most cells. In L. squamigera 2n = 27 (6 M + 11 st + 10 t) - a probable hybrid of L. straminea 2n = 16 (6 M + 10 t) and L. sprengerii 2n = 22 st - such cell analyses as 2 IV + 1 III + 5 II + 61 and 3 IV + 5 II + 5 I were common. The combined data demonstrated considerable homology between M types from one group of species and st types from the x = 11 group, confirming a fusion or fission relationship between them but raising the question of the evolutionary connection between the t and st chromosomes and the manner of their differentiation.
Addressing that problem, Kurita (1988b) favored the view that x = 6 was the original basic number for the genus and that fission produced complements of 2n = 22 t which, in the more northern species, were converted into sts by pericentric inversion.
In contrast, Inariyama (1951) had earlier proposed that the ancestral karyotype was 2n = 22 t, giving rise to the 2n = 12 (10 M + 2 t) by successive homozygous fusions. He proposed that x = 11 was the original basic number, a conclusion supported by the high frequency of x = 11 in other genera of the Amaryllidaceae. Here, too, it was supposed that t chromosomes were transformed to st chromosomes by some type of centric shift before or during the differentiation of the northern species.
Since the data can clearly accommodate both views, any decision between them must rest on matters of probability or general consideration of the evolution of karyotypes as discussed at the conclusion of this review.
Diploids in this genus are either 2n = 8 M (Crosa, 1972) or 2n = 10 (6 M + 4 t) (Nunez et al., 1972). Differing by a single homozygous fusion or fission, they give rise to other species by chromosome-doubling with or without hybridization between them (Nunez et al., 1974, 1990). Two are 2n = 16 M and are cytological autotetraploids. One is an amphidiploid, 2n = 18 (14 M + 4 t) between the two diploid types, and is a genomic tetraploid. Another, Nothoscordum inodorum, 2n = 19 (13 M + 6 T), which provides evidence for the homology of metacentric arms with t chromosomes, is an apomict once considered a diploid (Levan & Emsweller, 1938); it probably combines three 3 M + 2 t genomes with one of 4 M (Nunez et al, 1974). Nothoscordum bonariense, 2n = 26 (22 M + 4 t) with NF = 48, is a genomic allohexaploid. A diagram of possible origins based on the conclusions of Nunez et al. was given earlier by Jones (1978).
This genus provides a good example of the influence of fusion or fission in diploids subsequently changing basic number and confusing the polyploid status of hybrid species. Here again the direction of change is unclear. Though a single homozygous fission at the diploid level may be the more likely explanation, the large chromosomes of this genus, coupled with a basic number of x = 4 as compared with many x = 8 species in its allied genus Allium, raises the possibility that it could have originated by ancient fusions from 2n = 16 t ancestors.
The slipper orchids Paphiopedilum, Phragmipedium, and Cypripedium are primitive orchids with large chromosomes. Extensive multiple whole-arm rearrangements have been found differentiating species in the first two genera (Karasawa, 1978, 1979, 1982, 1986; Karasawa & Tanaka, 1980). Cypripedium, with a constant chromosome number of 2n = 20 but varying karyotype patterns, shows no evidence of this type of change. Cox and co-workers provide additional information on chromosome variation including measurements of total DNA amount (in Bennett & Leitch 1997, Cox et al., 1998) and review the data in the context of phylogeny elucidated by a parsimony analysis of nuclear ribosomal internal transcribed spacer (ITS) analysis from almost 100 species (Cox et al., 1997a). Although they revise the generic groupings of Paphiopedilum, the classification used by Karasawa is maintained in the following description of his cytological data.
In Paphiopedilum all species of subgen. Brachypetala and those in several sections of subgenera Polyantha and Paphiopedilum have similar karyotypes with 2n = 26 with metacentrics and submetacentrics only.
Species in other sections, including all but one species in subgen. Barbata, have chromosome numbers in excess of 2n = 26 (i.e., 2n = 28, 30, 32, 33, 34, 35, 36, 37, 38, 40, 41, and 42), the increase in chromosome number being due to replacements of M's and sm's by t's equivalent in number to their dissociated arms. There appears to be an increase in DNA amount with rising chromosome number (Cox et al., in press). Cribb (1987) attaches considerable importance to the geographic distribution of the species, many of those with higher chromosome number being endemics of restricted population size on Malaysian islands, some species known only as a single population that were possibly founded by every few individuals.
The South and Central American genus Phragmipedium has fewer species than Paphiopedilum. Fifteen of these have individuals with 2n = 18, 20, 21, 22, 28, 30, and 40 with some B chromosomes in two of the 2n = 20 species (Karasawa, 1980). Some of the differences in chromosome morphology seem attributable to undefined structural changes, but numerical change, as in Paphiopedilum, is due to whole-arm rearrangements. Two species with the lowest number 2n = 18 have only M or sm types of different sizes. Increases of 2n to 20, 22, 28, and 30 showed the correlated loss of bi-armed chromosomes and the increase in T's with a constant NF = 36. Two of the 22-chromosome species did not form part of this series, having a calculated NF of 34. All of the Phragmipedium species have lower genome size than Paphiopedilum and show a narrower range in 4C values (Cox & Abdelnour et al l.c.). The data is insufficient to indicate whether or not there is any significant change with chromosome number increase.
In a recent study of another allied monotypic genus, Mexipedium, Cox et al. (1997b) described M. xerophytica as 2n = 26 (20 M + 6 t) and NF = 46. There is no obvious structural relationship of this karyotype with any of those in the three companion genera.
The extremes of variation in Paphiopedilum and Phragmipedium could be seen as a fusion or fission, but the distribution of the species and knowledge of population sizes in the former at least strongly supports Karasawa's view that fission increase has occurred. The pressures produced by small population size and possible enforced inbreeding could be responsible both for the induction and selection of karyotype variants and the consequent expansion of genetic variability.
Cypripedium, supposedly the most primitive of the three, has presumably evolved its species karyotypes in different ways, resulting in, at one extreme, C. debile with only M chromosomes and, at the other, C. japonicum with 8 st's in its complement (Karasawa & Aoyama, 1986). Chromosome number is uniformly based on x = 10, with most species diploid 2n = 20. The 4C DNA amount for three species shows wide variation, from 86.2 to 129.5 pg (Cox et al., in press).
In the first period of cytological investigation, culminating in the survey of several genera by Marchant (1968), karyotype descriptions were based on small samples of taxa often of uncertain provenance. They showed predominance of species with 2n = 16 or 2n = 18 and karyotypes with varying degrees of asymmetry dominated by M and sm types of differing length. A pair of T's present in some genera was replaced in others by a pair of sm's with the shorter arm represented by a multisegmented accumulation of nucleolar organizing regions(NORs). This concentration of NORs was in contrast to their dispersal to terminal regions of many chromosomes where T's were present. Marchant suggested that this difference showed evolutionary progression caused by the accumulation of dispersed NORs onto a single chromosome pair. There was no evidence of fusion or fission in the family, though the description by Sax and Beal (1934) of acutely asymmetric karyotypes in Cycas 2n = 22 (12 M + 10 T) and Microcycas 2n = 26 (4 M + 22 T) led Marchant to hypothesize that these could represent the starting- or end-point of two series of progressive increase or decrease in chromosome number by such means. Recent descriptions of karyotypes of five species of Cycas in Japan, showing that the largest chromosomes are st and t types and the smallest M and sm (Kondo et al., 1995; Kokubugata & Kondo, 1996), do not provide support for that suggestion.
Positive evidence for the involvement of fusion or fission in the cycads has resulted from sampling natural populations in the genus Zamia in Mesoamerica. Norstog (1980, 1981) found some species to be 2n = 16 (12 M + 4 sm), supporting Marchant's findings, but found others 2n = 18 (10 M + 2 sm + 2 st + 4 T), 2n = 24 (4 M + 20 T) and 2n = 27 (1 M + 4 sm + 6 st + 16 T). Within Z. chigua (= Z. rozelii) variants with 2n = 22, 24, 25, and 26 were 4 M + 4 st + 2 st + 12 T at one extreme and 4 sm + 2 st + 20 T at the other, with several types existing in the same population. Despite some inconsistency in NF values - probably attributable to uncertainty in determining precise chromosome morphology and the possibility of arms being added by NOR translocations - the correlated changes in M and T frequencies are typical of a fusion or fission series. The studies of Vovides (1983) show the same relationship between 2n = 16 and 2n = 18 species.
Moretti and Sabato (1984) also discovered intraspecific variation in Mexican populations of Z. paucijuga: 2n = 23, 24, 25, 26, 27, or 28, the series beginning with 5 M + 2 sm + 8 st + 8 T and ending with 2 sm + 8 st + 18 T and showing overall progressive replacement of one M by two T's. High chromosome numbers with some intraspecific variation was later found in other species (Moretti, 1990a, 1990b; Moretti et al., 1991; Caputo et al., 1996). In these later papers the authors designate as T some chromosome types previously described as st's, in the belief that their obvious short-arms are better interpreted as enlarged centromeres. That change, though perplexing, does not affect the conclusion that chromosome number variation in the genus both within and between species is principally the consequence of fusion or fission.
Zamia loddigesii from seven localities in Yucatan occurred as 2n = 17, 24, 25, 26, 27 (Vovides & Olivares, 1996). Here, too, the changes in number seem attributable to fusion or fission, though the clarity of metacentric-telocentric relationships is partly obscured by other types of structural change.
Reviewing the chromosome data for Zamia, Moretti (1990a) points out that the karyotype diversity found in some species parallels the wide variability in phenotypic characters in the genus compared with the low phenotypic variability in the stable karyotypes of Ceratozamia 2n = 16 (12 M + 2 sm + 2 T) and Dioon 2n = 18 (8 M + 8 sm + 2 T). Although finding no unequivocal evidence to support a choice between the two alternative mechanisms for changing chromosome number, Caputo et al. (1996) see fission as the more likely in the light of supporting evidence from cladistic analysis and considerations of vicariance, dispersion, and extinction events. And although the changes in chromosome number were seen as the probable result of stressful environmental factors, these factors could not be specifically identified.
The gymnospermous family Podocarpaceae, sometimes referred to as "southern conifers," has an ancient history and relict distribution combined with extensive and extreme variation in chromosome number of its diploid species (Hair, 1963, 1966; Hair & Beuzenberg, 1958a, 1958b). In their first paper, Hair and Beuzenberg (1958a), characterizing chromosomes in somewhat general terms as "V" and "I" types, gave chromosome numbers and karyotype patterns for 53 species covering all genera of the family. Some additional data privately provided is given in Jones, 1978.
Hair and Beuzenberg found the following series of variants: 2n = 20 M, 2n = 22 (18 M + 4 t), 2n = 24 (16 M + 8 t), 2n = 26 (14 M + 12 t), 2n = 30 (10 M + 20 t), 2n = 34 (6 M + 28 t), 2n = 36 (4 M + 32 t), 2n = 38 (2 M + 36 t), and 2n = 40 t, all with NF = 40. Three species of Dacrydium - 2n = 18 (14 M + 4 t), 2n = 22 (10 M + 12 t), and 2n = 24 (8 M + 16 t) - fell into another group with NF = 32, and three Phyllocladus species were all 2n = 18 M, NF = 36.
The regular mathematical relationship between the number of M and t chromosomes indicates the involvement of fusion and or fission in producing the two extremes of complete metacentricity and complete telocentricity; t-M-t trivalents in a hybrid confirmed the relationships of related M and t's. In the few examples published - i.e., the idiograms of two Podocarpus species in Hair & Beuzenberg, 1958a, and four of Hair's camera lucida drawings of species. in three genera in Jones, 1978 - it is clear that there is considerable variation both within and between genera. At one extreme, metacentrics are of similar size, and if t or T chromosomes are present they equate in length with M arms. At the other extreme, the karyotype of Dacrydium laxifolium designated V chromosomes which seem to include M and sm types are amongst the smallest members of the complement, while the I chromosomes include t or T types of widely varying length but most being three to four times the length of any metacentric. Between these extremes there is in Podocarpus marked variation in length between V's and between the l's, some of the latter equating with t or T types and others with st types.
While acknowledging the paucity of adequate data due partly to the technical difficulties, Hair considered series of fusions as the prime source of variation reducing chromosome number from 2n = 40 to 2n = 20. Khoshoo (1962), in contrast, considered on principle that the highly asymmetrical karyotypes must be seen as extremely specialized and advanced, and consequently suggested fission as the more likely mechanism. The choice between these alternatives cannot yet be made. Since the extent of fusion/fission variation in the family has no known parallel in higher plants, the resolution of the problem could provide information of profound importance for assessing the role of both types of change. The data on chromosome constitution could be increased, but knowledge of the environmental factors affecting the survival of populations and the spread of species, coupled with phylogenetic relationships exposed by molecular means, would seem to be the more profitable.
In general, it is not possible to account for the differences of basic number x = 7, 8, 9, and 10 in the genus by whole-arm rearrangements. T and t chromosomes are rare, and where they do appear [as in Allium roseum, x = 8 (7 M + 1 t)] at diploid, tetraploid, and pentaploid levels, it is likely that the t's are the result of centric shift in sm chromosomes of similar length. The exceptions are species of the A. erdelii group found in the Negev (Kollman, 1970). There A. qasynense 2n = 14 M, A. erdeii 2n = 16 (12 M + 4 t), and A. negevense 2n = 20 (8 M + 12 t) show the Robertsonian relationship and suggest a fission series from 2n = 14 M. Allium zebadense 2n = 18 (14 M + 4 t) may also be a case of fission from 2n = 16 M.
Crocus has a very wide range of karyotypes and basic numbers with no indication of the involvement of fusion or fission in the evolution of its 60-70 species. The only exception is C. minimus, commonly 2n = 24 in Sardinia and Corsica but with three populations in Corsica lacking the typical chromosome number and containing only those with 2n = 25, 26, 27, 28, 29, and 30 in various combinations (Brighton, 1978). Because the population samples were small, it was impossible to know the precise extent of their chromosome variability. However, Brighton records that all of the increases in chromosome number were caused by the replacement of some M or sm chromosomes by telocentrics which seemed to equate to their dissociated arms. Seen against the background of the karyotype patterns in Crocus in general and that of the 2n = 24 members of the same species, one must agree with Brighton that successive centric fissions are responsible for the observed variation.
IV. Miscellaneous Cases
Table I lists genera in which basic number changes have been attributed to fusion or fission but generally without sufficient evidence to positively distinguish between them. Chaenactis is exceptional in that hybrids were produced between two x = 4 and their closely related x = 5 species. Meiotic pairing in the F1s showed that the x = 4 complements were the consequence of a single homozygous fusion between different pairs of chromosomes in the two species, respectively. That data and the rarity of x = 4 makes fusion an inevitable conclusion. Unlike the simple cases of fusion/fission described earlier, this case is complicated by various other interchanges.
In Viscum the change from x = 12 to x = 11 in V. fischeri has taken place in one genome for which the male plants are heterozygous. As females are homozygous for the x = 11 genome, the one with x = 12 never appears independently as a homozygote.
Not included in the table, because of imprecise evidence, is the case of Haplopappus (Compositae), where it is considered likely that at least one fusion was involved in the origin of H. gracilis 2n = 4 from H. ravenii 2n = 8 (Jackson, 1962).
V. Final Comment
The number of suspected cases of fusion and fission increases, but most of these are simple records of karyotype comparisons of small numbers of individuals unable to distinguish between them or add to knowledge of their significance. The general comments must then be made against a background of a dearth of investigational studies.
Robertsonian fusion has been demonstrated as a significant mechanism of karyotype evolution in several genera of the Tradescantieae where it can be regarded as a type of chromosome ortho-selection. In the T. commelinoides complex, successive fusions, in combination with polyploidy, have converted a 2n = 14 t complement to one with 2n = 14 M, and conceivably may also have produced a similarly confusing situation in Gibasis matudae and G. pulchella. Although it is likely that most of the fusions there would have arisen in ancestral diploids, and there is some evidence to support it, no diploid species has yet been found with more than a single homozygous difference. That raises the possibility that diploids with maximum fusions were supplanted by the polyploids that arose from them. On the other hand, although the studies of the Tradescantieae have been unusually comprehensive, the population samples (as in most investigations of species) are few and are from restricted localities. There is a need for much more extensive sampling if the chromosome evolution in the subfamily is to be exposed and understood.
There is sound evidence for the origin of two diploid species of Chaenactis by different fusions from a single ancestor. Despite Kurita's reservations, it also seems probable that the 2n = 12 karyotypes in Lycoris represent, as Inariyama and Koyama suggested, the culmination of a series of fusions from an ancestral complement of 2n = 22 t. The question of the relationship between that constitution and those with 2n = 22 st and the significance of the difference remains to be solved and could be an inviting subject for FISH exploration.
Apart from these cases, there is no sound evidence for fusion. The remarkable diversity in the Podocarpaceae certainly involves either fusion or fission or a combination of both, and its solution could add significantly to the general problem.
[TABULAR DATA FOR TABLE I OMITTED]
Fission seems the best explanation for the rare increases in chromosome number in Allium, Crocus, and Crinum. In Nothoscordum, too, it may account for the 2n = 10 species, though it may be premature to exclude fusion as the explanation for the large chromosomes of the genus.
The increases of chromosome number in slipper orchids can be attributed only to series of fissions. In Zamia, too, that seems a reasonable explanation for the inter- and intraspecific variation. In both cases it has been suggested that karyotype changes occur in populations under stress - small populations produced by few founders in the case of the orchids and possible erosion of populations in Zamia. Consequent enforced inbreeding could promote chromosome breakage with increase in genetic variability advantageous for survival and/or for colonization of new habitats.
If the fissions in the above examples are, indeed, the consequence of population pressures, then they are secondary modifications of karyotypes without direct bearing on the evolution of the lower chromosome numbers from which they arose. If we wish, then, to speculate on the more ancient evolution of the latter, it could be that fusions played a part in their origin. In the Cycadales, for example, some have suggested that the highly asymmetric complements of Cycas and Microcycas are relicts of an ancestral constitution. It seems premature to dismiss that possibility, particularly bearing in mind the extremely asymmetric 2n = 46 t complement of Welwitschia (Khoshoo & Ajuha, 1963).
In considering the bearing of both fusion and fission on karyotype evolution, there may have been cycles of change - one producing stable karyotypes by fusion and another repatterning by fission to be followed again by the stabilizing influence of fusion - but such events are too remote for easy access, at least by classical cytological techniques. Molecular systematics and molecular cytology may be able to expose what did take place in ancient lineages. Meanwhile, it seems unwise to interpret directions of chromosome evolution merely on the basis of the symmetry-asymmetry dogma. We are able to observe only one episode of karyotype change, and, depending on its nature, symmetry may be either primitive or derived.
This review has focused attention on only two related types of chromosome change, both of which are found to occur in a relatively small but diverse group of plants. Although both are tempting candidates for producing changes in basic number in many plant groups, it cannot be concluded that either has played a role in karyotype evolution comparable to that found in animal species. Putting aside the possibility that their involvement has been obscured, it is well known that inversions, interchanges, tandem fusions, losses, and duplications make significant contributions to karyotype change. Only the continuing exploration of karyotype diversity can lead to a better understanding of the relative roles that each has played in the evolutionary process. The armoury of modern techniques now provides opportunities of an exceptional sort.
VI. Literature Cited
Bennett, M. D. & I. J. Leitch. 1997. Nuclear DNA amounts in Angiosperms - 583 new estimates. Ann. Bot. 80: 169-196.
Bidau, C. J. 1990. The complex Robertsonian system of Dichroplus pratensis (Melanoplinae: Acrididae). II. Effects of the fusion polymorphisms on chiasma frequency and distribution. Heredity 64: 145-159.
-----. 1993. Causes of chiasma repatterning due to centric fusions. Rev. Brasil. Genet. 16: 283-296.
Brandham, P. E. 1976. The frequency of spontaneous structural change. Pp. 77-87 in K. Jones & P. E. Brandham (eds.), Current chromosome research. North-Holland Publishing, New York.
-----. 1983, Evolution in a stable chromosome system. Pp. 251-260 in P. E. Brandham & M. D. Bennett (eds.), Kew Chromosome Conference II. Allen & Unwin, Boston.
Brighton, C. A. 1978. Telocentric chromosomes in Corsican Crocus L. (Iridaceae). Pl. Syst. Evol. 129: 299-314.
Caputo, P., S. Cozzolino, L. Gaudio, A. Moretti & D. W. Stevenson. 1996. Karyology and phylogeny of some Mesoamerican species of Zamia. Amer. J. Bot. 83: 1513-1520.
Cox, A. V., S. T. Bennett, A. S. Parokonny, A. Kenton, M. A. Callimassia & M. D. Bennett. 1993. Comparison of plant telomere locations using a PCR-generated synthetic probe. Ann. Bot. 72: 239-247.
-----, D. G. Abdelnour, M.D. Bennett & I. J. Leitch. 1998. Genome size and karyotype evolution in the slipper orchids (Cypripedioideae: Orchidaceae). Amer. J. Bot. 85: 681-687.
-----, A. M. Pridgeon, V. A. Albert & M. Chase. 1997a. Phylogenetics of the slipper orchids (Cypripedioideae: Orchidaceae). Nuclear rDNA - its sequences. Pl. Syst. Evol. 208: 197-223.
-----, ----- & M. A. T. Johnson. 1997b. Cytological characterization of Mexipedium xerophyticum (Cypripedioideae: Orchidaceae). Lindleyana 12: 162-165.
Cribb, P. J. 1987. The genus Paphiopedilum. Collingridge, London.
Crosa, O. 1972. Estudios cariologia en el genero Nothoscordum (Liliaceae). Bol. Fac. Agr. Uruguay 122: 3-8.
Darlington, C. D. 1956. Chromosome botany. Allen & Unwin, London.
Gibby, M., S. Hinnah, E. Marais & F. Albers. 1996. Cytological variation and evolution within Pelargonium sect. Hoarea (Geraniaceae). Pl. Syst. Evol. 203: 111-142.
Goldblatt, P. 1979. Chromosome cytology and karyotype change in Galaxia (Iridaceae). Pl. Syst. Evol. 133: 61-69.
-----. 1980. Uneven diploid numbers and complex heterozygosity in Homeria (Tridaceae). Syst. Bot. 5: 337-340.
Hair, J. B. 1963. Cytogeographical relationships of the southern podocarps. Pp. 401-414 in J. L. Gressitt et al. (eds.), Pacific basin biogeography. Bishop Museum Press, Honolulu.
-----. 1966. Biosystematics of the New Zealand flora, 1945-1964. New Zealand J. Bot. 4: 559-595.
----- & E. J. Beuzenberg. 1958a. Chromosomal evolution in the Podocarpaceae. Nature 181: 1584-1586.
----- & -----. 1958b. Contributions to the chromosome atlas of the New Zealand flora 1. New Zealand J. Sci. 1958: 617-628.
Hunt, D. R. 1980. Sections and series in Tradescantia. American Commelinaceae IX. Kew Bull. 35: 437-442.
Inariyama, S. 1932. Cytological studies of the genus Lycoris 1. Conjugation of chromosomes in meiosis in L. albiflora Koidz. Bot. Mag. Tokyo 46: 426-434.
Jackson, R. C. 1962. Interspecific hybridization in Haplopappus and its bearing on chromosome evolution in the Blepharodon section. Amer. J. Bot. 49: 119-132.
John, B. & M. Freeman. 1975. Causes and consequences of Robertsonian exchange. Chromosoma (Berlin) 52: 123-136.
Jones, K. 1974. Chromosome evolution by Robertsonian translocation in Gibasis. Chromosoma (Berlin) 45: 353-368.
-----. 1977. The role of Robertsonian change in karyotype evolution in higher plants. Pp. 121-129 in A. de la Chapelle & M. Sorsa (eds.), Chromosomes today: Proceedings of the 6th International Chromosome Conference. Oliver & Boyd, Edinburgh.
-----. 1978. Aspects of chromosome evolution in higher plants. Adv. Bot. Res. 6: 120-194.
-----. 1990. Robertsonian change in allies of Zebrina (Commelinaceae). Pl. Syst. Evol. 172: 263-271.
----- & S. Bhattarai. 1981. Contributions to the cytotaxonomy of Commelinaceae. Gibasis linearis and its allies. Bot. J. Linn. Soc. 83: 141-156.
----- & A. Kenton. 1984. Mechanisms of chromosome change in the evolution of the tribe Tradescantieae (Commelinaceae). Pp. 143-168 in A. Sharma & A. K. Sharma (eds.), Chromosomes in evolution of eukaryotic groups. CRC Press, Florida.
----- & J. B. Smith. 1967. Chromosome evolution in the genus Crinum. Caryologia 20: 163-179.
-----, D. Papes & D. R. Hunt. 1975. Contributions to the cytotaxonomy of the Commelinaeeae II. Further observations on Gibasis geniculata and its allies. Bot. J. Linn. Soc. 71: 145-166.
-----, A. Kenton & D. R. Hunt. 1981. Contributions to the cytotaxonomy of the Commelinaceae. Chromosome evolution in Tradescantia sect. Cymbispatha. Bot. J. Linn. Soc. 83: 157-188.
Karasawa, K. 1978. Karyomorphological studies on the intraspecific variation of Paphiopedilum insigne. La Chromosomo II-9: 233-255.
-----. 1979. Karyomorphological studies in Paphiopedilum, Orchidaceae. Bull. Hiroshima Bot. Gard. 2: 1-149.
-----. 1980. Karyomorphological studies in Phragmipedium, Orchidaceae. Bull. Hiroshima Bot. Gard. 3: 1-49.
-----. 1982. Karyomorphological studies on four species of Paphiopedilum, Orchidaceae. Bull. Hiroshima Bot. Gard. 5: 70-79.
-----. 1986. Karyomorphological studies on nine taxa of Paphiopedilum. Bull. Hiroshima Bot. Gard. 8: 23-42.
----- & M. Aoyama. 1986. Karyomorphological studies in Cypripedium in Japan and Formosa. Bull. Hiroshima Bot. Gard. 8: 1-22.
----- & R. Tanaka. 1980. C-banding study on centrio fission in the chromosomes of Paphiopedilum. Cytologia 45: 97-102.
Kenton, A. 1981a. Chromosome evolution in the Gibasis linearis alliance. The Robertsonian differentiation of G. venustula and G. speciosa. Chromosoma (Berlin) 84: 291-304.
-----. 1981b. A Robertsonian relationship in the chromosomes of two species of Hydrocleys (Butomaceae sens. lat.). Kew Bull. 36: 487-492.
-----. 1982. Chromosome evolution in Mexican species of Gibasis (Commelinaceae). Ph.D. thesis, Reading University, Reading, U.K.
-----. 1984. Robertsonian differentiation and preferential pairing revealed in species and F1 hybrids of the Gibasis linearis group (Commelinaceae). Pl. Syst. Evol. 144: 221-240.
-----, A. Davies & K. Jones. 1987. Identification of Renner complexes and duplications in permanent hybrids of Gibasis pulchella (Commelinaceae). Chromosoma (Berlin) 95: 424-434.
-----, S. J. Owens & D. Langton. 1988. The origin of ring-formation and self-compatibility in Gibasis pulchella (Commelinaceae). Pp. 75-84 in P. E. Brandham (ed.), Kew Chromosome Conference III. Her Majesty's Stationery Office, London.
Khoshoo, T. N. 1962. Cytogenetical evolution in the gymnosperms. Proc. Summer Sch. Bot. Darjeeling: 119-135.
----- & M. R. Ajuha. 1963. The chromosomes and relationships of Welwitschia mirabilis. Chromosoma (Berlin) 14: 522-533.
King, M. 1993. Species evolution - the role of chromosome change. Cambridge University Press, Cambridge.
Kokubugata, G. & K. Kondo. 1996. Differential fluorescent-banding in chromosomes of four species of Cycas (Cycadaceae). Bot. J. Lima. Soc. 120: 51-55.
Kollman, F. 1970. Karyotypes of three Allium species of the erdelii group. Caryologia 23: 647-655.
Kondo, K., G. Kokubugata, M. Hizume, R. Tanaka & T. Satake. 1995. A karyomorphological study of five species and one variety of Cycas. Cytologia 60: 141-147.
Koyama, M. 1962. Meiosis in Lycoris albiflora. Ann. Rep. Doshisha Women's College 12: 9-12.
-----. 1978. Chromosome pairing in the genus Lycoris II. Ann. Rep. Doshisha Women's College 29: 272-282.
Kurita, S. 1986. Variation and evolution in the karyotype of Lycoris I. General karyomorphological characteristics of the genus. Cytologia 51: 803-815.
-----. 1987a. Variation and evolution in the karyotype of Lycoris II. Karyotype analysis of ten taxa among which seven are native in China. Cytologia 52: 19-40.
-----. 1987b. Variation and evolution in the karyotype of Lycoris III. Intraspecific variation in the karyotype of L. traubii Hayward. Cytologia 52: 117-128.
-----. 1987c. Variation and evolution in the karyotype of Lycoris IV. Intraspecific variation in the karyotype of L. radiata (L'Herit.) Herb. and the origin of this triploid species. Cytologia 52: 137-149.
-----. 1988a. Variation and evolution in the karyotype of Lycoris VI. Intrapopulational and/or intraspecific variation in the karyotype of L. sanguinea Maxim. vat. kiushiana Makino and vat. koreana (Nakai) Koyama. Cytologia 53: 307-321.
-----. 1988b. Variation and evolution in the karyotype of Lycoris VII. Modes of karyotype alteration within species and probable trend of karyotype evolution in the genus. Cytologia 53: 323-335.
Kyhos, D. W. 1965. The independent aneuploid origin of two species of Chaenactis (Compositae) from a common ancestor. Evolution 19: 26-43.
Levan, A. & S. Emsweller. 1938. Structural hybridity in Nothoscordum fragrans. J. Hered. 29: 291-294.
-----, K. Fredga & A. A. Sandberg. 1964. Nomenclature for centromeric position on chromosomes. Hereditas 52: 201-220.
Marchant, C. J. 1968. Chromosome patterns and nuclear phenomena in the cycad families Stangeriaceae and Zamiaceae. Chromosoma (Berlin) 24: 100-134.
Mattsson, O. 1971. Cytological observations within the genus Zebrina. Bot. Tidsskr. 66: 189-227.
Maxted, N., M. A. Callimassia & M. D. Bennett. 1991. Cytotaxonomic studies of eastern Mediterranean Vicia species (Leguminosae). Pl. Syst. Evol. 177: 221-234.
Moretti, A. 1990a. Karyotypic data on North and Central American Zamiaceae (Cycadales) and their phylogenetic implications. Amer. J. Bot. 77: 1016-1029.
-----. 1990b. Cytotaxonomy of cycads. Mere. New York Bot. Gard. 57: 114-122.
----- & S. Sabato. 1984. Karyotype evolution by centromeric fission in Zamia (Cycadales). Pl. Syst. Evol. 146: 215-233.
-----, P. Caputo, L. Gaudio & D. W. Stevenson. 1991. Intraspecific chromosome variation in Zamia (Zamiaceae, Cycadales). Caryologia 44: 1-10.
Moscone, A., M. Lambrou, A. T. Hunziker & F. Ehrendorfer. 1993. Giemsa C-banded karyotypes in Capsicum (Solanaceae). Pl. Syst. Evol. 186: 213-229.
Norstog, K. 1980. Chromosome numbers in Zamia. Caryologia 33: 419-428.
-----. 1981. Karyotypes of Zamia chigua (Cycadales). Caryologia 34: 255-260.
Nunez, O. 1990. Evolucion cariotipica en el genero Nothoscordum. Acad. Nac. Ca. Ex. Fis. Nat., Buenos Aires. Monogr. 5: 55-61.
-----, N. Frayssinet & R. Rodriguez. 1972. Los cromosomas de Nothoscordum Kunth (Liliaceae). Darwinia 17: 243-245.
-----, -----, ----- & K. Jones. 1974. Cytogenetic studies in the genus Nothoscordum Kunth. I. The N. inodorum polyploid complex. Caryologia 23: 235-254.
Parker, J. S. 1987. Increased chiasma frequency as a result of chromosome rearrangement. Heredity 58: 87-94.
-----, A. S. Wilby & S. Taylor. 1988. Chromosome stability and instability in plants. Pp. 131-140 in P. E. Brandham (ed.), Kew Chromosome Conference III. Her Majesty's Stationery Office, London.
Redi, C. A., S. Garagna & E. Cappana. 1990. Nature's experiment with in situ hybridization? A hypothesis for the mechanism of Rb fusion. J. Evol. Biol. 3: 133-137.
Rousi, A. 1961. Cytotaxonomical studies on Vicia cracca L. and V. tenuifolia Roth. Hereditas 47: 81-110.
Sax, K. & J. M. Beale. 1934. Chromosomes of the Cycadales. J. Arnold Arb. 15: 255-258.
Schubert, I. & R. Reiger. 1990. Alteration by centric fission of the diploid number in Vicia faba L. Genetica 81: 67-69.
Stone, D. E. & J. L. Freeman. 1968. Cytotaxonomy of Illicium floridanum and I. parviflorum (Illiceaceae). J. Arnold Arb. 49: 41-51.
Strid, A. 1968. Stable telocentric chromosomes formed by spontaneous misdivision in Nigella doerfleri. Bot. Not. 121: 153-164.
Vovides, A. P. 1983. Systematic studies on the Mexican Zamiaceae. I. Chromosome numbers and karyotypes. Amer. J. Bot. 70: 1002-1006.
----- & M. Olivares. 1996. Karyotype polymorphism in the cycad Zamia loddigesii (Zamiaceae) of the Yucatan peninsula, Mexico. Bot. J. Linn. Soc. 120: 77-83.
Weins, D. & B. A. Barlow. 1975. Permanent translocation heterozygosity and sex determination in East African mistletoes. Science 187: 1208-1209.
Werner, J. E., R. S. Kota, B. S. Gill & T. R. Endo. 1992. Distribution of telomeric repeats and their role in the healing of broken chromosome ends in wheat. Genome 35: 844-848.
White, M. J. D. 1945. Animal cytology and evolution. Ed. 1. Cambridge University Press, Cambridge.
-----. 1973. Animal cytology and evolution. Ed. 3. Cambridge University Press, Cambridge.
|Printer friendly Cite/link Email Feedback|
|Publication:||The Botanical Review|
|Date:||Jul 1, 1998|
|Previous Article:||Orbicules in angiosperms: morphology, function, distribution, and relation with tapetum types.|
|Next Article:||Taxaceae: the genera and cultivates species.|