Meiotic chromosome behaviour in Cenchrus ciliaris ( Poaceae : Pani-coideae )

A basic chromosome number of x = 9 has been confirmed for Cenchrus ciliaris L. Polyploidy is common and levels vary from tetraploid to hexaploid. Aneuploidv is reported for a single specimen, where two chromosomes of a single genome were lost. Various meiotic irregularities were observed. The highest incidence of meiotic abnormalities was observed in the pentaploid specimens. This was attributed to their uneven polyploid level All specimens varied from segmental alloploid to alloploid.


INTRODUCTION
The classification of the genus Cenchrus L. is com plex (De Lisle 1963) and it is sometimes difficult to describe species on gross morphological characters alone (Chapman 1990).Meiotic chromosome behaviour can make a vast contribution to the classification process.
The aim of the present study was to use meiotic chro mosome behaviour in C. ciliaris in order to deduce the polyploid origin oi the specimens studied, as well as the chromosomal variation within and among populations of C. ciliaris in South Alrica.

MATERIALS AND METHODS
Specimens were collected and fixed in the veld.Voucher herbarium specimens are housed in the Geo Potts Herbarium.Department of Botany and Genetics, University of the Orange Free State, Bloemfontein (BLFU) (Table 1).
Slides were prepared for meiotic analysis (Visser & Spies 1994).A minimum of 20 cells of each of the fol lowing stadia were studied: diakinesis, metaphases I and II, anaphases I and II and telophases I and II.The fol lowing were recorded: gametic chromosome numbers, the number of univalents, laggards and micronuclei dur ing metaphase I, anaphase I and II.and telophase I and II, respectively.The number of chiasmata per cell was inferred from the chromosome configurations observed during diakinesis or metaphase I.
The genomic constitution of some of the tetraploid plants was calculated, based on the models proposed by Kimber & Alonso (1981).Expected chromosome config urations for each of the proposed models (4:0, 3:1, 2:1:1 and 2:2) were calculated and compared with those observed.The average sum of squares (SS) between the expected and the observed values were calculated.The relative affinity between the pairs of genomes was expressed as an x-value.An x-value of 0.5 indicated a close affinity between corresponding (hom ologous) genomes, whereas non-homologous genomes were rep resented by an x-value of 1.The model with the lowest SS-value provided the best 'fit' for that particular speci men.The specific model was then considered to repre sent the genomic constitution of the specimen.
A wide range of meiotic abnormalities were observed for Cenchrus ciliaris.These abnormalities included the presence of univalents during metaphase I. chromosome and chromatid laggards during anaphases I and II respec tively, uneven segregation ot chrom osom es during anaphase I. anaphase I bridges and micronuclei during telophases I and II.Two additional meiotic abnormalities were observed, namely precocious disjunction of chro mosomes into chromatids during anaphase I and the absence of cytokinesis at the end of telophase I.
of the three chromosome abnormalities will be represen tative of all the tetraploid specimens.This process was not repeated for the pentaploid and hexaploid specimens, due to the limited number of specimens available.
The average numbers of univalents observed in the specimens investigated were 1.5, 1.9, 4.0 and 2.8 respec tively for the n = 17, n = 18, n = 45/2, and n = 27 speci mens.The variation in the number of univalents is indi cated in Table 2.The highest number of univalents (0-18) was observed for Spies 5230 (n = 18) (Figure 2B, C).
The univalents present were usually situated near the equator (Figure 2A -D 2).The variation in numbers is indicated in Table 2.The highest number of micronuclei (11) was observed for Spies 5210 (n = 45/2) (Figure 4B).The size of these micronuclei varied (Figure 4B, D).
Genome interpretation was performed on three tetraploid specimens {Spies 5215, 5240 and 5649).These analyses revealed that the 2:2 model (Kimber & Alonso 1981) fitted the specimens to the greatest degree, with xvalues of 1 or approximately 1 (Table 3).5A-C).The average number of cells per specimen which contained bridges, varied from nought to a maximum of 71.4% (Spies 5652) (Table 2).The acentric fragment could most often be observed (Figure 5A, B and hexaploids were observed at much lower frequencies (9.2% and 6.6% respectively).

Anaphase bridges were observed in 26 specimens of C. ciliaris (Table 2) (Figure
Polyploidy is prominent in the plant kingdom (Stebbins 1982).Polyploid levels for C. ciliaris, taking the published aneuploid chromosome numbers into account, include diploidy, triploidy, tetraploidy, pentaploidy, hexaploidy and nanoploidy.References for these a -  The meiotic chromosome behaviour of the polyploid specimens varied from being normal to highly irregular, depending on the number of genomes present and their homology.Polyploids with an even number of genomes are influenced less dramatically during meiosis than those with an uneven number of genomes.Uneven poly ploid levels have more meiotic abnormalities, due to the presence of an uneven number of sets of chromosomes which com plicates chromosome pairing.Genome homology also plays an important role in the normality of meiosis.It can affect chromosome pairing to such an extent that from univalents to multivalents are formed.
In order to attest the proposed basic chromosome number of x = 9 for C. ciliaris, a comparison was made between the observed and the expected meiotic chromo some behaviour of each of the polyploid levels studied.The presence of univalents during metaphase I was reg ularly observed.A variation in the number of univalents present occurred within and among the polyploid levels (Figure 2A-D ).The highest average numbers of univa lents were observed in the pentaploid specimens (4.0), whereas for the hexaploid specimens, they were 2.8 (Table 2).Approximately similar average numbers of univaients were observed for the aneuploid and tetraploid specimens (1.5 versus 1.9).
The increased averages of univalents present during metaphase I for the pentaploids and the hexaploids were expected, as a higher incidence of meiotic irregularities is closely associated with uneven and higher polyploid levels (Stebbins 1950).The variation in the number of univalents observed within each of these polyploid lev els, indicates chromosomal differences among the speci mens.
The number of univalents present during metaphase I corresponds with the numbers of chromosome laggards and micronuclei observed (Figure 1).The data suggest that the univalents observed during metaphase I were mostly lagging during anaphase I (Figure 3A-E), and led to the formation of micronuclei during telophase I.The average number of laggards was the highest for the pentaploid specimens (9.2 laggards per cell), followed by the hexaploid and aneuploid specimens (3.3 and 2.9 respec tively) (Table 2).The number of chromosome and chro matid laggards varied among the specimens of the vari ous polyploid levels, emphasizing genetic differences among specimens within each of the polyploid levels.
The average number o f micronuclei also varied among the specimens of each polyploid level (Table 2).The micronuclei were not incorporated into the daughter nuclei at the time of cell division (Figure 4A -D ).The highest average number was observed in the pentaploid specimens (4.1), followed by the hexaploid (3.8), tetraploid (1.1) and aneuploid (0.9) specimens respec tively.The highest number and greatest variation of micronuclei were observed in the pentaploid specimens (Figure 4B, C) (Table 2).
The highest average numbers of univalents, chromo some laggards and micronuclei have been observed in the pentaploid specimens (Table 2).This polyploid level was followed by the hexaploids, whereas for the tetraploids and the single aneuploid specimen, the level was approximately similar.The high incidence of meiot ic abnormalities in the pentaploid specimens was expect ed, due to their uneven polyploid levels.An average number of 9.2 chromosomes lagged during anaphase I and was representative of an entire genome.A basic chromosome number of x = 9 for C. ciliaris, due to the presence of a fifth genome lagging during anaphase I, is hereby confirmed.Genome variation in the tetraploid specimens was confirmed during this study.Meiosis was normal in some specimens (Spies 5522), whereas for others, it was high ly irregular (Spies 5230) (Table 2).The average numbers o f univalents present and chromosome laggards observed, were almost similar (1.9 and 1.7, respectively).These averages suggest two univalents observed during metaphase 1, lagging during anaphase I, and finally form ing a single micronucleus (1.1 per cell) (Table 2).3).An x-value of 1 inferred two dis  3).This specimen was classified as a segmental alloploid, based on the lower x-value and the occasional presence of quadrivalents (for example AAA'A').

Genome interpretation of three tetraploid specimens
The specimens' genomic constitutions were attested by their meiotic chromosome behaviour.The specimens differed in respect o f their meiotic behaviour.Quadrivalents were observed in Spies 5215 and 5649, whereas for Spies 5240, univalents and bivalents were observed.The number of quadrivalents observed in the two specimens varied.One to three quadrivalents were observed for Spies 5649, whereas for Spies 5215, an infrequent quadrivalent was observed.
Spies 5240 was confirmed as an alloploid (for exam ple AABB), based on the presence of bivalents only, observed during metaphase I.The presence of an occa sional quadrivalent in cells of Spies 5215 and one to three quadrivalents per cell in Spies 5649, justify seg mental alloploidy (for example AAA'A') for both these specimens.However, the presence of genes, controlling homoeologous chromosome pairing in Cenchrus, should be studied before these genomic constitutions can be accepted.
The two alloploid specimens were collected in the Eastern Cape, whereas Spies 5649 was collected in the Free State (Table 1).Taking their different geographical localities into consideration, it is suggested that the Eastern Cape specimens represent a genetically different group (or hybrid swarm) from those in the Free State.Therefore, chromosomal variation in these three tetraploid specimens confirm the presence of genetic variability in C. ciliaris in South Africa.
The highest incidences of meiotic abnormalities were recorded in the pentaploid specimens (Table 2).These abnormalities were mostly the result of an uneven poly ploid level.The average number of univalents observed, was approximately half of the average number of chromo some laggards observed (4.0 versus 9.2) (Table 2).There was an increase in the average number of laggards observed (Table 2).This could have been due to amphitelic orientation and equational distribution of the univalents observed during metaphase I, as various chromatid lag gards were observed during anaphase I (Figure 3C).These anaphase I laggards were mostly included in more than one micronucleus per cell (Figure 4B, C), as an average frequency of 4.1 micronuclei per cell has been observed.The average number of laggards observed for this poly ploid level, was representative of a complete genome lag ging during anaphase I.It is suggested that the genomic constitution of the pentaploid specimens includes one unrelated genome.This suggestion is based on the average number of univalents present, the absence of trivalents and the high average number of laggards observed during anaphase I.A genomic constitution of, for example AAA'A'A", is proposed for this polyploid level.Segmental alloploidy is justified by the high and the low occurrence of bivalents and quadrivalents respectively, observed dur ing prophase and metaphase I. Chromosomal variation is evident in this polyploid level, for meiotic chromosome behaviour varied among the specimens studied (Table 2).
The average number of univalents present, the chro mosome and chromatid laggards and the micronuclei observed for the hexaploid specimens (2.8, 3.3 and 3.8 respectively), were relatively low when compared to that of the pentaploid specimens (Table 2).A genomic con stitution of.for example A A A A A 'A ", is proposed for the hexaploids.Segmental alloploidy, based on the abun dant bivalent and occasional quadrivalent chromosome configurations found (Figure 2D), is proposed.The pres ence of quadrivalents indicates a degree of homology between the A and A' genomes.
Hybridisation among plant individuals is usually characterised by various changes in chromosome struc ture (Darlington 1937;Dobzhansky 1941).For C. cil iaris, these changes include the presence of paracentric inversions.These inversions were mostly observed in the tetraploid specimens (Figure 5A -C ).The higher occur rence of these inversions could be due to the high num ber of tetraploid specimens which were cytogenetically studied.The highest average number of cells containing anaphase 1 bridges was observed in Spies 5652 (71.4%) (Figure 5A) (Table 2).Different paracentric inversions were found among the specimens, as the chromosome fragments differed in size.
Occasional or recurrent hybridisation and the com plete local breakdown of reproductive isolation between sympatric species result in the production of hybrid swarms.These swarms include the whole range of genet ic variability of the parental species.This scenario could be representative of C. ciliaris, for a wide range of chro mosomal.morphological and genetic variation is evident in this species.

CONCLUSIONS
With the aid of meiotic analyses, a basic chromosome number of x = 9 has been confirmed for Cenchrus cil iaris.Polyploidy is common and varies from tetraploid to hexaploid.Aneuploidy was observed in a single spec imen.It is suggested that this specimen is the result of loss aneuploidy from two chromosomes of a single genome.
Various meiotic irregularities were observed for this species.The highest incidences of meiotic abnormalities were observed for the pentaploid specimens.This was attributed to their uneven polyploid and chromosome number.
The chromosome abnorm alities observed during meiosis were an indication of genomic relationships.These relationships varied among the specimens and the polyploid levels.Segmental alloploidy was suggested for the aneuploid specimen, whereas for the tetraploid specimens, it varied from segmental alloploidy to allo ploidy.A genomic constitution of alloploidy and seg mental alloploidy is suggested for the pentaploid and hexaploid levels respectively.The nature of the genom ic relationships indicated the presence of hybridisation.Hybridisation in C. ciliaris was confirmed by the chro mosomal variation observed among specimens in each of the polyploid levels.

TABLE 2
.-Meiotic chromosome behaviour of Cenchrus ciliaris specimens showing voucher specimen no.; gametic chromosome no.(n); aver age frequency of univalents (I); frequency of chromosome laggards; percentage of cells studied containing anaphase I bridges; frequency of micronuclei during telophase I.All ranges are included in brackets

TABLE 3 .
Kimber & Alonso (1981)of tetraploid Cenchrus ciliaris specimens, according to models ofKimber & Alonso (1981): voucher nos, chiasma frequencies, relative x-values and the appropriate average sum of squares for each of the various models in brackets (M)