zyxwvuts zyxwvuts zy Botanical Journal ofthr Ltnnean Society, 81: 301-325. With 36 figures December 1980 Cytotaxonomy of Commelinaceae: chromosome numbers of some African and Asiatic species zyxwvu R. B. FADEN Department of Botany, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560, U . S . A . AND Y. SUDA Department of Biology, Faculty $Education, Iwate University, Ueda, Morioka 020,Japan zyxwv zyxw zyx zyxwvutsr Acceptedfor publication June 19RO Chromosome counts are reported for 32 taxa (31 species and 1 subspecies) belonging to 10 genera of Commelinaceae from seven African and Asiatic countries. Counts for 13 species and 1 subspecies are recorded for the first time. Published chromosome numbers for A n h i c o p s i s and Polyspatha are confirmed. It is suggested that Pdisota, Pollia and Stanfieldidla each has a single basic number (x = 20, 16 and I I , respectively).The known cytological diversity in Floscopa is extended. The third continental African species of Coleolrype is found to have the same chromosome number (2n= 36) as the other two. The preponderance of the basic number x = 15 in Commlim is supported. The uncommon basic number x = 13 is reported in four taxa of Cyanotis together with karyotypic differences. The basic number x = 6 is found in a second species of Murdannia. Karyotypic data in addition to chromosome numbers are presented for 24 of the 32 taxa investigated. Karyotypes are found to be useful in assessing relationships in the family, and evolutionary trends in the karyotype are noted. KEY WORDS:-Africa-Asia-basic Commelinaceae- karyotypes numbers-chromosome evolution-chromosome numbers- CONTENTS Introduction . . Materials and methods . . . . Results AnthericopstJ . Coleotrype . . Commelina . . Cyanotis . . Floscopa . . Murdannia . . Palisola . . . Pollia . . . Polyspatha . . Stanjeldiella . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 1 0024-4074/80/080301 + 25$02.00/0 . ,286 ,286 ,287 ,289 ,291 ,291 ,293 ,295 ,297 ,297 ,299 ,299 ,299 zyxwvuts 0 1980 The Linnean Society of London 302 zyxwvutsrq zy zy zyxwvu R. B. FADEN AND Y. SUDA . . . . . . . . . . . Discussion Similarities and differences in karyotypes . Primitive and advanced species within genera The derivation of one genus from another . Uniformity of karyotypes within genera Evolutionary direction in aneuploid series Acknowledgements . . . . . . . . References . . . . . . . . . . . Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ,301 ,301 ,302 ,303 ,303 ,304 ,305 ,306 .307 zyxwvuts zy INTRODUCTION The Commelinaceae has been a favourite subject for cytological study. As a result of numerous investigations, chromosome counts have been published for 42 of the 49 genera and for approximately 37%of the species. Counts are not available for Aetheolirion, Elasis, Phaeosphuerion (excluding Commelinopsis), Pseudoparis, Sauvallea, Spatholirion and Triceratella, none of which comprises more than two species. Despite the abundance of data, there have been few attempts at a cytological review of the family. (Goldblatt, 1980; and Jones and Jopling, 1972, may be mentioned as exceptions.)A major reason for this is the difficulty in using the published information. Many generic concepts are still in a state of flux, and counts for some species have been published under several names, a source of confusion which is compounded by the absence of recent monographs for nearly all genera. Furthermore, the variety of counts recorded for some taxa, particularly ones from India and West Africa, has necessitated using some data with great caution. Although the best approach to this problem would be the monograph in which cytological studies are combined with morphological and other investigations, careful cytological surveys of diverse genera and species can clarify our knowledge of the family. In addition to chromosome number, other karyotypic data, such as chromosome size and centromere position, can be useful for the assessment of generic or specific relationships. MATERIALS AND METHODS Mitotic and meiotic counts were made from plants grown (from field collected cuttings or seeds) mainly at the Missouri Botanical Garden, St. Louis. Generally, root tips for mitotic counts were collected in late morning or early afternoon. Root tips were pretreated with a saturated solution of paradichlorobenzene for approximately three hours (except Stanzeldiella which was not pretreated), then b e d in modified Carnoy’s fluid (4 chloroform: 3 absolute ethanol: 1 glacial acetic acid, v/v) for at least 2-3 h, and finally transferred to 70%ethanol and stored in a deep freeze. For examination, the root tips were hydrolyzed with 1.2 N HC1 for about 40 min, then immersed in acetic-orcein and kept at room temperature for at least 24-48 h (sometimes up to 7 - 10 days). Buds for meiotic counts were fixed and stored in modified Carnoy’s fluid and kept in a deep freeze. They were stained with acetic-orceinwhen examined. Temporary slides were made by the squash method and quickly examined under the microscope. Slides with good figures were made permanent by a modified McClintock method (McClintock, 1929) for further study (6.Faden, 1975). Photomicrographs were taken from these slides with a Zeiss Universal Research Microscope equipped with an Automatic Photomicrographic Camera. Drawings were made from the photomicrographs while comparing them with the CYTOTAXONOMY O F COMMELINACEAE 303 original figures under the microscope. Voucher slides are in the possession of one of us (R. B. F.). Voucher specimens either in the form oforiginal field collections or pressings of the cultivated plants, or both, have been deposited in various institutions, one of which is indicated in each case in the Appendix. zyxwv zyxwvuts zyxwvu zy RESULTS Our results are listed in Tables 1 and 2. Counts were considered exact only if at least three unambiguously countable figures were obtained. Generally, considerably more than three figures were counted, but tallies of numbers of counts were not kept. The karyotypic data included in Table 2 are based on all permanized figures, notjust on the one illustrated. The number of taxa counted in each genus is as follows, with the number of taxa counted for the first time in parentheses: Anthericopszs- 1 (0); ColeotTpe- 1 ( 1); Commelinu- 12 (5-4 species and 1 subspecies); Cyunotis-5 (4); Floscopu- 1 (0); Murdunnzu-3 (2); zyxwvu zyxwvutsrqponml Table 1. Chromosome counts for Commelinaceae .. Species count reported here n Anthericopsis sepalora Coleolrype bruecknerana Commelinahenghalensis ~ - _ Previous reports 2n n 14 - 36 22 _ - 22 Ganguly (19461, Mitra & Datta (1967) Kanlmdth\ & Rdo \1961a), Panigrahi & Kammathv( 1962) Bhattacharva (19751, Fuji5hima (19691, Fukumoto (19641, Panigrahi & Karnmath? (1964),Sharma(1955), Shetty & Subramanyam (1961, 1962) Malik ( 1961) Zaman & Ahmed ( 1972 I Lewis (1964) Chimpharnba( 19731 Jones &Jopling( 1972) He m ( 1968 11 22 11+(0-2BI 11 11,22 22+(0-4B) 33 c. 24 - Jones&Jopling( 1972) - - - 14 11 - Commelina bracteosu Commelinaconges fa Commelina dlfjusa (sometimes I-ecorded as Commeha nud@ora L. 1 2n none - - Reference - - - 30 30 30 - - - - - zyxw zy 44 44,66 - 2 2 , 2 8 , 44, 56,66 - 28,56 30 c. 68 none 28 28 28,56 - 28 - - - 56 28,30, 60 2 8 , 30, 56, 58,60 30 15,30 - Morton 11967 Anderson & Sax 1 1936) Morton(19561 Harve) (1966) Darlington (1929al Morton(1956, 1967 H s u ( l 9 7 I ) , Morton (1956 Sharma &Sharma ( 19581 Anderson & Sax ( 1 936) Darlington (1929a) Bhattachana (1975) Morton (1967) Chimphamba( 19731,Jones&Jopling - zyxwvutsr R. B. FADEN A N D Y. SUDA 304 zyxwvutsrq zyxwvut Table 1. Chromosome counts for Commelinaceae (continued) Count reported here Species n Commelina dcffusa (cont.) Commelinafoliacea subsp.foliacea Commelinafoliacea subsp. ‘A’ Commelina lat$lia Commelina manosperma Commelina thamarii Cummelina trilobosperma Commelina rambesica Commelina sp. ‘E’ of Faden (1974) Cyanotisfoecunda Cyantotis nyctitropa Cyanotispaludosa Cyanotis sp. ‘A’ Cyanotis sp. ‘B’ Floscopa africana subsp. petrophila Murdannia clarheana Murdannia simplex (sometimesrecorded as Aneilema sinicum Ker-Gawl.) Murdannia zeylanica Palisota barteri Palisota hirsuta 2n Previous reports n Reference 2n 30 120 30,35,42 (1972);Lewis, Stripling&Ross ( 1962);Lewis, Suda & Oliver ( 1967 1 ; Simmonds (1954) Kammathy & Rao ( 1965);Lewis (1964); Panigrahi&Kammathy(1962, 1964); Raghavan & Rao ( 1961) Kammathy&Rao(l96la) Jones&Jopling(l972) Sharma(1955) 30 20 20,22 56 56,60 Jones&Jopling (1972) Morton (1956) Morton (1967) Morton (1956) Morton ( 1967) 26 c. 56 Chimphamba(l973) Morton (1967) none 30 30 22 60 30 56 30 26 26 22 26 78 none zyxwvuts none none 18 16 24 40 - none none none none none 20 - - - 40 - - 40 40 40 none Palisota orientalis Pollia condensata 40 32 none Pollia secundlflora 32 40 32 - 40 32 - zy Morton (1967) Heitz (1969,F.afncana, undifferentiated) Hsu ( 1967) Heitz (1969);Hsu( 197 1);Lewis (1964) Chimphamba (1973);Jones &Jopling (1972);Morton(1966, 1967) Kammathy & Rao ( 196lb) Panigrahi & Kammathy ( 1961) Morton ( 1967) MiPge(1960) Jones &Jopling ( 197 2); Morton ( 1967 Jones&Jopling(l972);Mangenot& Mangenot(l958, 1962);Morton (1967) Hsu (1972, as P. sorzogomsis (E. Mey.) Steud.) Kammathy&Rao(1965) zyxwvutsrq zyxwvutsr - Pollca thyrslflora Polyspatha paniculata Stanfreldiella impefo rata var. impeforata Stanfreldrella olrgantha 32 28 32 none - - 28 Jones &Jopling ( 1972);Morton (1967) Heitz (1968) 22 22 44 24 Morton (1967) Jones& Jopling( 1972) Morton(l967) 22 CYTOTAXONOMY OF COMMELINACEAE 305 Table 2. Karyotypic data for some species of Commelinaceae zyx zy zyxwvutsrqpon zyx zyxwvutsrqponml zyxw zyxwvutsrqpon Taxon Anthericopris sepalosa 2n Mean chromosome size Within-complement variation Chromosome Chromosome size’ typest 14 medium bimodal medium: 4m+2sm small: 2smt+6sm 36 22 22 56 30 30 30+(0-4B) 30 30 very large large medium large medium medium large medium medium various various various various various various various similar various m, sm m, sm m, sm m, sm, st m, sm m, sm m, sm, st m, sm m, sm Cyanotispaludosa 22 large various C.foecunda C . nyctitropa C . sp. ‘A’ C. sp. ‘B’ 26 26 26 78 large large large large similar. similar similar various large: 2 m medium: 2m+2mt small: I6 t St m, st Floscopa africana subsp.fetrophila 16 medium bimodal Murdannia clarkeana M . simplex M . zeylanica 24 40 40 small verysmall small various similar bimodal m, sm Palisota hirsuta 40 large various Pollia condensata P . secundtflora 32 32 small small various various Polyspatha paniculata 28 large Stanjeldiella oligantha 22 verysmall Cokotrype bruecknerana Commelina benghalensis C . macrosperma C . zambesica C . bracteosa C. dtjiusa C .foliacea subsp .,foliacea C . latfolia C . sp. ‘E’ ofFaden (1974: Number ot pairs of satellite chromosomes Figure I, 2 1 at least I 2 at least 1 at leayt 2 3 at least 1 at least 1 2 at least 1 35, 36 4 3 10, I I 14, 15 7, 8 12, 13 9 5, 6 I 20 1 I I at least I 16, 1 7 19 18 21, 22 none 23, 24 at least 3 25 26 27, 28 m, sm at least 1 29 m, sm, s t m, sm, s t at least 1 at least 1 31 30 various m, sm, st none various small : 2 m very small : 20 sm St St medium: 2m+6sm small: 4m+4sm 32 1 33, 34 zyx zyxw * Similar: chromosomes all approximately the same size; bimodal: chromosomes of two size classes; various: chromosomes variable in size but not bimodal. t The following abbreviations are used: m, metacentric; sm, submetacentric; st, subtelocentric; t, telocentric. (Superscript‘t’ refers to satellite chromosomes.) Palisota- 3 ( 1 ) ; Pollia-3 ( 1 ) ; PoLyspatha- 1 (0); Stanfieldiella-2 (0). The counts obtained for Commelina congesta, Floscopa africana subsp. petrophila and Stanjieldiella oliguntha differ from previous counts reported for these taxa. Anthericopsis The 2n = 14 count obtained for this monotypic, eastern African genus confirms those reported by Jones & Jopling (1972). Our figures (Figs 1 , 2 ) also accord with that of these authors. The low basic number x = 7 among Old World genera in the family has previously been noted (Jones 8c Jopling, 1972). zyxwvutsrq zyxwvutsrq zyxwvutsrqp zyxwvutsrqpon zyxw zyxwvuts Figiiirs 1-9. Figs 1, 2. Antherimpsic sepalosa (slide number 75/549): 2n= 14. Fig. 3. Commelina miicrv\prmfl (75/270): 2 n = 2 2 . Fig. 4. Commlina benghalensis (751413):2n=22. Figs 5, 6 . Commelina sp. ‘E’ ( 7 5 1 5 6 2 ) :2 n = 3 0 . Figs 7 , 8. Commelina d@sa (751157):2n=30. Fig. 9.Commelina latfofolia(75/442): Zn=30. zyxwv zyxwv zyxwvutsr zyxwvuts zyxw zy zyxwv zyxwvu CYTOTAXONOMY O F COMMELINACEAE 307 Coleotrype The 2n = 36 count which we report for the East African endemic C. bruecknerana (Figs 35, 36) agrees with those recorded for the two other continental African species of the genus, the southern African C. natalensis C. B. Clarke (Anderson & Sax, 1936; Darlington, 1929a; Guervin 8c Le Coq, 1966; Heitz, 1968; Jones 8c Jopling, 1972)and the western African C.laurentii K. Schum. (Morton, 1967).The only different count in the genus is 2n = 42, recorded for an unnamed Madagascan species (Jones & Jopling, 1972). Coleotrype has the largest chromosomes of any genus examined in this investigation and, along with the closely related genus Forrestia, has the largest chromosomes among Old World genera of Commelinaceae (Jones & Jopling, 1972). A basic number x = 9 was suggested for Coleotrype by Jones & Jopling (1972) because of the occurrence of that basic number in Forrestia (based on the count 2n = 18 for F. tenuis (C. B. Clarke) Benth. (Morton, 1967)).However, x = 6 is also possible, and the Madagascan species could be a hexaploid based on x = 7. An alternative hypothesis is that x = 18 is the basic number for the continental African species of Coleotrype. In that case the Madagascan plant would likely be a diploid based on x = 2 1. The occurrence of high basic numbers in such genera as Palisota (x = 20) and Dichorisandra and its relatives (x = 19) indicates that this hypothesis merits further consideration. The sole meiotic investigation in the genus is that of Heitz (1968) whose photograph of metaphase I in C. natalensis shows only bivalent pairing, suggesting that the plant might have been a diploid with the basic number x = 18 (or possibly an allotetraploid). Certainly, further studies of chromosome pairing behaviour during meiosis would be useful, particularly of the taxa from Madagascar, which is the centre of taxonomic diversity in the genus. Commelina We report counts for 12 taxa ( 1 1 species and an undescribed subspecies of C. foliacea), five for the first time. The counts confirm the preponderance of the basic number x = 15 in the genus. However, we also find x = 14 in C.zambesica and x = 1 1 in C. benghalensis and C. macrosperma. Our count for C.zambesica, 2n = 56 (Figs 10, 1 l), confirms the approximate count of Morton (1967) for this species but differs from the 2n = 26 count published by Chimphamba (1973). Although this species is quite variable, we suspect that Chimphamba’s plant may not have been conspecific with ours. Our counts of 2n = 22 for C. macrosperma (Fig. 3) and 2n = 60 for C. thomasii, agree with the higher numbers of Morton (1967)but not with his earlier counts (Morton, 1956). Similarly, we find 2 n = 30 for C. congesta from Ghana, whereas Morton (1956, repeated in Morton, 1967)reported 2n = 28, also from Ghanaian material of this species. As Jones & Jopling (1972) have pointed out, the counts reported by Morton ( 1956)for 19 taxa of Commlina are extraordinary in that none is based on x = 15. In fact every one except that for C. macrosperma is based on x = 14. When 10 of these same taxa were recounted by him (Morton, 1967), five of them then included counts based on x = 15 (as well as additional counts in some cases). Two of the remaining five species also had different basic numbers in the later paper: C. benghalensis (x= 11, in addition to the earlier x= 14) and C. macrosperma EL zyxwvut zyxwvutsr zyxwv zyxwvu zy zyxw CYTOTAXONOMY OF COMMELINACEAE 309 (cf.Table 1). We agree with Lewis ( 1964) that it does seem peculiar that infraspecific aneuploidy in Commelina would appear to be almost confined to the West African species. We also concur with Jones &Jopling ( 1972)that considerably more studies of these species are necessary. In addition to the basic numbers x = 11, 14 and 15 reported here and in the literature, the only basic numbers which we consider reliably recorded for Commelina are x = 13 (Chimphamba, 1973; Lewis, 1964) and x = 12 (Kammathy 8c Rao, 1961a). Curiously, all of the taxa recorded with x = 13 are from southcentral Africa (Malawi to eastern Zaire) which is the major centre of taxonomic diversity for the genus. The single species with x = 12 is C. attenuata Vahl from India. The basic numbers x = 10 and 16 have also been reported in Commelina, but not convincingly. The former is based solely on Morton’s count for C. macrosperma (Morton, 1956) which neither he (Morton, 1967)nor we could confirm. The latter is reported by Lewis ( 1964) as having come from the literature, but we are unable to trace the source. Of the genera investigated here, we find the chromosomes of Commelina to be medium-sized to relatively large. However, Anderson & Sax ( 1936) and Jones & J o p h g ( 1972) list them as small for the family as a whole. The chromosomes are mostly meta- and submetacentrics, and, except in C.macrosperma and C.zambesica, there is relatively little variation in size within the complement. zyxwvut Cyanotis According to the literature, chromosome counts have been reported for approximately half of the 50 species of Cyanotis (including Amischophacelus). Basic numbers range from x = 8 to x = 14, with x = 12 most frequent, x = 8 reported once (Shetty & Subramanyam, 1962) and needing confirmation, x = 9 unrecorded and x = 14 noted only in some plants of C. ‘somaliensis’ (cf. below). The basic number x = 11, which we report here for C.paludosa (Fig. 20), is rare in the genus. This basic number has been noted previously: in a plant determined as C. arachnoidea C. B. Clarke-probably misidentified, based on the description given-from India (Kammathy & Rao, 196lb); in some African populations of C. barbata D. Don (Lewis, 1964; Lewis & Tadesse, 1964), and in the Ethiopian endemic C.polyrrhiza Hassk. (Lewis & Tadesse, 1964).Whether o r not these species are related remains to be demonstrated, but it may be noted that Asiatic populations of C. barbata-sometimes determined as C.uaga Lour. -are based on x = 12 (cf. Rao, Raghavan & Kammathy, 1972, for summary; and Bhattacharya, 1975). The remaining four taxa reported here have a basic number of x = 13 which is also uncommon in Cyanotis. This number has been recorded in the African C. foecunda (Lewis, 1964; confirmed here, Figs 16, 17) and C. speciosa (L. fil.) Hassk.-synonym C. nodzjiora (Lam.)Kunth-(Chimphamba, 1973; Guervin & Le Coq, 1966; Heitz, 1968; Jones 8c Kukkonen, 1971; Lewis, 1964); in some populations of the Asiatic C. thwaitesii Hassk. (Raghavan & Rao, 1965; Sundararaghavan & Rao, 1965) and C. uillosa (Spreng.) Schult. fil. (Jones 8c Kukkonen, 1971; Raghavan & Rao, 1965; Rao, 1970); in C. cristata (L.) Schult. fil. (Jones & Jopling, 1972-correct1 determined?) and in cultivated material determined as C. somaliensis C . B. C arke (Guervin & Le Coq, 1966; Jones & Kukkonen, 197 1 ; Matsuura & Sutb, 1935). zyxwvu r zyxwvutsrq zyxwvutsr zyxwvu zyxwvuts Figures 14-20. Figs 14, 15. Comrmlitla bracteosa (slide number 751567): 2n=SO. Figs 16, 17. Cyanotis foecunda (751587):2n=26. Fig. 18.CyanotiC sp. 'A' (7616):2n=26. Fig. 19. Cyamtis nyctitropa (751595): 2n=26. Fig. 20.CyanotiCpdudora(751505):2n=22. zyx zyxwvuts zyxw zyx CYTOTAXONOMY OF COMMELINACEAE 311 zyx Cyanotis nyctitropa, for which we report 2n = 26, is an endemic of the Arabian Peninsula which has been confused with the Ethiopian C. polyrrhiza. Cyanotis nyctitropa also shows morphological similarity to the south Indian and Sri Lankan species C. pilosa Schult. El., but may be distinguished from both C. pilosa and C. polyrrhiza by the absence of stalked inflorescences. In the present study we find that it also differs from them in basic chromosome number: x = 1 1 for C. polyrrhiza (cf. above); x = 12 for C. pilosa (Jones & Jopling, 1972; Jones & Kukkonen, 197 1 ; Kammathy & Rao, 196lb1, and x = 13 for C.nyctitropa (Fig. 19). Cyanotis nyctitropa was originally considered by the first author to be closely related to C. foecunda because both species possessed strictly sessile inflorescences and contiguous ranges. The present study has confirmed that they also have the same chromosome number. However, C. nyctitropa has at least two pairs of metacentric chromosomes (Fig. 19),while C.foecunda has only subtelocentrics (Figs 16, 17). Whether o r not these karyotypic differences are constant requires further investigation. The two remaining Cyanotis taxa, C. sp. ‘A’ and C.sp. ‘B’ belong to what the first author calls the ‘C. foecunda complex.’ In the herbarium of the Royal Botanic Gardens, Kew, specimens are filed under C.foecunda, as is the holotype of C. somaliensis. In other institutions these unnamed taxa are sometimes segregated out as C. somaliensis. Both C. sp. ‘A’ and C. sp. ‘B’ are endemic to East Africa and possess sessile, axillary inflorescences and shoots with indeterminate growth. Ultimately, they may be treated as subspecies despite their different chromosome numbers, Their affinity to one another and to C.foecunda is attested to by their basic number x = 13 and by the absence of metacentric chromosomes (Figs 18,21, 22). The identity of the cultivated plant sold under the name %. somaliensis’ is still uncertain. Of the plants studied here, it most closely resembles in morphology the hexaploid which we are calling C.sp. ‘B’. However, according to published counts, C. ‘somaliensis’ is always diploid with a basic number of either x = 13 (cf. above) o r x = 14 (Anderson & Sax, 1936; Darlington, 1929a, 1929b, 1932). Whenever the karyotype is indicated, three or four pairs of metacentric chromosomes are mentioned or illustrated (Anderson & Sax, 1936; Darlington, 1929a, 1932; Guervin & Le Coq, 1966: fig. 38; Jones & Kukkonen, 197 1: fig. 2a; Matsuura & Sut6, 1935).As yet there is no known wild plantwhich has this karyotype. Floscopa The 2n = 16 count which we report for F. africana subsp. petrophila from Uganda (Figs 23, 24) differs from the 2n = 18 count of Morton (1967) for material of the same subspecies apparently from Ghana. It agrees, however, with the n = 8 count reported by Heitz (1969) from an unspecified subspecies of F . africana of undesignated origin. The only additional count for this species is that of Morton (1967) who recorded 2n = 36 for F. africana subsp. africana. Further study is required in order to determine whether o r not aneuploidy occurs in this species at the diploid level. Floscopa appears to be cytologically diverse. There are basic numbers of x = 6, 8 and 9, diploids, tetraploids and hexaploids, apparent autopolyploidy and perhaps infraspecific aneuploidy among the five species counted thus far. The bimodal complement which we illustrate has also been noted by Jones & Jopling (1972)in F. scandens Lour. -- z zy 12 zyxwvutsrq zyxwvutsrq zyxwvut Figurrb 21-24. Figs 21, 22. Cyanolrs sp. 'B' (slide number 75/555): 2n=78. Arrow denotes broken c11ronios01iie.Figs 23, 24. Floscopa africana subsp. petrophila (75/547):2n= 16. zyx zyxwv zyxwvu zy zyxw zy zyx CYTOTAXONOMY OF COMMELINACEAE M urdannia 313 There are counts in the literature for about 20 of the approximately 50 species in the genus. Unfortunately, many of the counts need confirmation. Nevertheless, it can be stated that the basic numbers x = 6, x = 9, x = 10 and x = 1 1 occur, and that x = 10 predominates. The report ofk = 21 and 2n = 42 for M. japonica (Thunb.) Faden (as Murdannia data (Vahl) Bruckn., Panigrahi 8c Kammathy, 1963) may indicate the additional basic number x = 7 . We report the chromosome number of M . zeylanica, 2n = 40 (Figs 27, 281, for the first time and find that this species has a different chromosome number, but the same basic number ( x = 10) as three previously counted, related species: M . divergens (C. B. Clarke) Bruckn. (n = 30, Panigrahi & Kammathy, 1961); M . esculenta (C. B. Clarke) Rao 8c Kammathy ( n = 10, Kammathy & Rao, 196lb); M . hookeri (C. B. Clarke) Bruckn. ( n = 10, Kammathy & Rao, 1965). Our count 2n = 40 for M . simplex confirms all previous counts from African material of this species (Chimphamba, 1973; Heitz, 1969 (as Aneilemasinicum);Jones &Jopling, 1972; Lewis, 1964; Morton, 1966, 1967).InAsia this species is cytologically more variable, and diploids, tetraploids, hexaploids and octoploids have been reported (Hsu, 1967, 197 1 ; Kammathy 8c Rao, 196lb; Panigrahi & Kammathy, 196 1 ; Raghavan 8c Rao, 196 1 ; Shetty 8c Subramanyam, 1961, 1962). Murdannia clarkeana, whose chromosome number 2n = 24 we report for the first time, belongs to a group of species which, unlike the genus as a whole, is principally African. The sole, related species which has been counted is M . semiteres (Dalz.) Santapau (including M . juncoides (Wight) Rao & Kammathy) which is recorded as n = 6 and n = 12 (Kammathy8cRao, 1961b, 1964, 1965). These are the only species in the genus with the basic number x = 6, which gives cytological evidence of a relationship first suggested on the basis of morphological characters (Brenan, 1952).The counts n = 7 , n = 10 and n = 20, also reported for M. semiteres by Raghavan & Rao ( 196 11, should be disregarded, as they could not be reconfirmed by these authors in later investigations (Rao et al., 1972). The chromosomes in all species are small. In M . reylanica there is a bimodal complement with approximately equal numbers of chromosomes of two size classes. Palisota The counts for the three species presented here support the conclusion ofJones & Jopling (1972) that Palisota has a single basic number x = 20. The 2n = 32 count reported by Miege ( 1960) for P. hirsuta was probably based on wrongly determined material (Pollia species ?) because all subsequent counts for this species have been 2n = 40 (Jones 8c Jopling, 1972; Morton, 1967; the present study, Fig. 29). The number n = 36 for P. bracteosa C. B. Clarke (Heitz, 1968) needs confirmation; all other counts for this species have been based on x = 20 (Anderson & Sax, 1936; Jones & Jopling, 1972; Morton, 1967). Our 2n = 40 count for C. barteri confirms that of Morton (1967).Our report of the same somatic chromosome number for the Tanzanian endemic P. orientalis agrees with that of Morton ( 1967)for the closely related species P. ambigua (P. Beauv.) C. B. Clarke from western Africa. The chromosomes of Palisota are relatively large. 21 zyxwvu zy zyxwvutsr zyx zyxwv CYTOTAXONOMY OF COMMELINACEAE Pollia 315 Including the three species counted in this investigation, chromosome numbers of 2n = 32 have now been reported for 8 of the 1 7 species of Pollia (including Aclisia E. Meyer). There is probably a single basic number, x = 16, and the three anomalous counts in the literature, 2n = 38 for P.japonica (Mitsukuri, 1947),n = 10 for P . secundzjora (Hsu, 1972) and 2n= 10 for P . subumbellata C. B. Clarke (Darlington, 19381, should be treated as having been based on wrongly determined material, because each of these species has also been counted as n = 16 or 2n = 32 (Fujishima, 1970a, 1970b, for P . japonica; Kammathy & Rao, 1965, and the present paper (Fig. 301, for P . secundiJora; Kammathy & Rao, 1965, for P. subumbellata). Our counts of 2n = 32 for P. condensata from Ghana (Fig. 3 1 ) and Tanzania confirm those of Mangenot & Mangenot (1958, 1962) and Morton (1967). Cytologically there are no significant differences among the species counted in Pollia sensu strictu (P. hasskarlii, P. japonica, P. thyrisflora) and those which would be assignable to Aclisia on the basis of androecial morphology (P. condensata, P. crispata, P. macrophylla, P . secundzjiora, P. subumbellata). The cytological uniformity supports the maintenance of Pollia senm lato for all these species (Faden, 1975). All Pollia species for which information is available have small chromosomes, much smaller than those of species of Palisota which was placed in the same generic “Group” by Brenan ( 1966) because of their indehiscent fruits. The cytological evidence does not indicate any close affinity between these genera and, along with morphological and anatomical data, it suggests that the indehiscent fruit arose separately in these genera through convergent evolution (Faden, 19 7 5). zyx Polyspath The 2n = 28 count for P. paniculata presented here agrees with all previous reports for this species and genus (Heitz, 1968; Jones & Jopling, 1972; Morton, 1967). Except for the absence of satellite chromosomes, our figure (Fig. 32) accords well with that of Jones & Jopling (1972). N o multivalent pairing was reported in P. paniculata by Heitz (19681, so we are in agreement with Jones & Jopling (1972) that the basic number is presumably x = 14, not x = 7 . Stanjeldiella Our experience in Stanjieldiella accords with that ofJones & Jopling (1972) and Morton ( 19671, the only previous workers on the cytology of the genus, that it has the smallest chromosomes in the family. Pretreatment proved to be unnecessary. Our 2n = 22 count for S. impeforata agrees with that of Morton (1967)and has the same basic number as the tetraploid plant of this species recorded by Jones & Jopling ( 1972). However, our 2n = 22 count for S. olzgantha (Figs33,341 differs from the 2n = 24 (based on sectioned material) for that species (Morton, 1967). Our counts indicate a single basic number x = 1 1 . We find, as did Morton ( 1967), that in all species one pair of chromosomes in the complement is considerably larger than the rest. 21‘ zyxwvutsrqpo zyxwvu Figures 31-36. Fig. 31. Pollia condensata (slide number 75/465): 2n = 32. Fig. 32. Polyspathapaniculata ( 7 5 / 6 3 9 ) :2n = 28. Figs 33,34. Stanfieldiella oligantha ( 7 5 / 4 5 9 ) :2n = 22. Figs 35,36. Coleotype bruechnerana(751385):2 n = 3 6 . zyxwvuts zyx zyx CYTOTAXONOMY OF COMMELINACEAE DISCUSSION 317 Although we have not worked out karyotypes in detail, we have included karyotypic data for most of the taxa investigated here (Table 2). The chromosome sizes given are for comparative purposes within this study; they do not necessarily agree with terms used in other works, such as Anderson 8c Sax ( 1936) and Jones & Jopling (1972). In fact, all of the chromosome complements which we treat in the size ranges ‘very small to small’ and ‘medium to large’ (Table 3) are in Jones & Jopling’s ‘small’category, while our ‘very large’ corresponds to their ‘medium’. (‘Large’ chromosomes, as used by Jones & Jopling, occur only in New World genera). We believe that we have added some refinement to their system, but the usefulness of this further division, if any, will have to be determined from future studies of more genera and species. The purpose of examining karyotypes even at this superficial level is to provide a basis for comparing species and genera. The significance of karyotypic similarities and differences for the assessment of relationships has never been discussed for the Commelinaceae as a whole. In general, we find that such data must be used with caution because of the frequency of aneuploidy and polyploidy within many genera in the family and because of the absence, for nearly all genera, of recent monographs. Thus, it is difficult to know whether a comparison is being made between the karyotypes of advanced species of one genus with karyotypes of less highly derived species of another genus. The following guidelines are recommended when comparisons are made between the karyotypes of different genera in the Commelinaceae. Similarities of karyotypes, particularly of distinctive ones, are better indicators of relationship than dissimilarities are of non-relationship. Only primitive species of different genera should be compared unless, (a)one genus is believed to have been derived from the other, or (b) karyotypes are uniform within one or both genera. In the absence of monographs which might identify primitive and advanced species within a genus, we believe that the following generalizations are useful: in an aneuploid series of basic numbers, evolution has generally gone in the direction of reduction; an accompanying trend has been an increase in chromosome size. Similarities and diflerences in karyotjpes We believe that the common possession of distinctive karyotypes by different genera strongly suggests a close relationship between them. The striking similarities in the karyotypes of Cochliostema, Dichorisandra, Geogenanthus and Siderasis, all of which have very large chromosomes and the unusual basic number Table 3. Chromosome size and basic numbers in the genera studied Very small to small Medium to large Very large zy zyxwvutsrqpon zyxwvutsrq zyxwvu zyxw 318 R. B. FADEN AND Y. SUDA zyxwvu 19, have been pointed out by Jones 8c Jopling (1972). Morphological similarities as well as mainly Amazonian distributions for these genera have been stressed by Hunt (1971). We are in agreement with Jones 8c Jopling (1972)that the karyotypic similarity among these genera is a significant indication of affinity. Similarly, Jones 8c Jopling ( 1972) have noted marked karyotypic similarity between the Old World genera Forrestia and Coleotrype. The genera are further alike in anatomical characters (Tomlinson, 1966)and in habit, inflorescence,floral and seed morphologies, Here again we concur that karyotypic likeness denotes true relationship. Hunt (1975) used the karyotypic identity of Setcreasea species and morphologically similar Tradescantia species as an important criterion for placing the former genus in synonymy under Tradescantia. The karyotype present in both, very large metacentric chromosomes and the basic number x = 6, seems to be restricted to these genera within the family and this restriction strongly supports their affinity. In contrast, we find that karyotypic dissimilarities are not necessarily indicative of non-affinity. For example, the genus Amischophaxelus was separated from Cyanotis by Rao 8c Kammathy ( 1966)partly on the basis of karyotype. Although karyotypic differences are present, Jones 8c Jopling ( 1972)have concluded, and we agree, that “the variation [in karyotypes] known in Cyunotis is sufficiently extensive to accommodate the karyotypes of C. axillaris and C . cucullata [the two species which comprise Amischophacelusl without embarrassment”. Furthermore, we find that there are no absolute differences in inflorescence and bracteole morphology (despite the claim to the contrary by Rao 8c Kammathy, 19661, nor in floral or seed morphology. The only morphological character which does set these two species apart from the rest of Cyanotis is the shape of the capsule apex, which we feel is hardly a sufficient difference to merit generic rank. Therefore we are in agreement with Brenan ( 1966)that C. axillaris and C . cucullatu should not be segregated from Cyanotis, despite the karyotypic differences. From the above examples one further generalizationmay be made, namely that karyotypic characters strongly reinforce other kinds of data, especially morphological data, but that when karyotypic information is at variance it should not be used to override important morphological characters in deciding generic limits, and perhaps not even species limits. However, when the morphological evidence is ambiguous, karyotypic data may assume much greater significance. This is particularly true for the tribe Tradescantieae (sensu Rohweder, 1956) in which only fine morphological distinctions separate the genera. x = zyx Primitive and advanced species within genera When comparing karyotypes of species belonging to different genera in an effort to assess generic affinities, it is important to compare only primitive species, except under special circumstances mentioned below. In Aneilema five basic numbers occur, x = 9, 10, 13, 15 and 16 (Faden, 1975).Correlativemorphological and anatomical data indicate that species with x = 16 are least advanced, while those with x = 9 are most advanced. When karyotype comparisons are made with other genera, only the species of section Aneilema (the most primitive in the genus, all species having the basic number x = 16)should be used because of‘thediversity in karyotype morphology within Aneilema. When this is done, a notable CYTOTAXONOMY OF COMMELINACEAE zy zyx 319 resemblance between the karyotypes of these species and those of Pollia is apparent. Morphological and anatomical data also indicate a relationship between the two genera (Faden, 1975). Thus karyotypic similarities between these genera support a close affinity between them. zyxwvu zyxwv zyxwvut zyxwvu zyxwv zyxwv The derivation of one genus from another When one genus is believed to have given rise to another on the basis of data other than cytology (to avoid a circular argument), the karyotype of the derivative genus, or those of the most primitive taxa therein, should be compared with that of the species (of the ancestral genus), whether primitive or advanced, which is thought to be most closely related to it. If the karyotypes are found to be similar, then the cytological information strongly supports the putative relationship. If, on the other hand, the karyotypes prove to be dissimilar, then the karyotype of the derived genus should also be considered derived, unless the presumed relationship between the genera was based on weak evidence. In the latter case the karyotypic data might provide a basis for the reassessment of the conjectured relationship. The genus Cyanotis is almost certainly derived from Belosynapsis. Within the Commelinaceae the two genera possess distinctive vegetative hair types (Tomlinson, 1966) and seed morphologies, as well as numerous other characters in common. In all morphological attributes in which they differ and for which evolutionary direction can be inferred, Cyanotis is more advanced. Within Cyanotis the most primitive species may possibly be found among C. uillosa (Spreng.)Schult. fil. and its relatives. Morphologically, these taxa most closely resemble B . kewensis Hassk. among Belosynapsis species. When the karyotypes of these species are compared (cf.Jones 8c Kukkonen, 197 I ) , it is found that that ofB. kewensis shows a basic number x = 13 with small, mostly metacentric chromosomes. Cyanotis villosa proves to be the only Indian species in the genus which regularly (at least in some populations) has a basic number x = 13. Its chromosomes are larger than those of B . kewensis and are mostly subtelocentrics (Jones & Kukkonen, 197 1 : fig. 2e). The karyotypic data are consistent with Cyanotis having been derived from Belosynapsis, suggesting that subtelocentric chromosomes have evolved from metacentrics in this lineage. Unqormity of karyotypes within genera I t is sometimes possible to infer non-affinity between genera without knowing the primitive species and without having monographs of one or both genera. When more than one-third of the species in a genus have been reliably counted and the cytological evidence points to a single basic number and karyotype for the genus, then it is reasonable to anticipate the same basic number and karyotype in the remaining, uncounted taxa, provided that none is morphologically eccentric. Surprises may certainly occur, and the chromosome numbers and karyotypes of the outstanding species should be examined. However, until such data are available, some logical inferences about the karyotype of the genus as a whole may still be made. By so doing, we deduce single basic numbers for Palisota ( x = 201, Pollia (x = 16) and Stanjieldiella (x = 1 1 ) . Furthermore, we find that the karyotypic 320 zyxwvu zyxwvut zyxw R . B. FADEN AND Y. SUDA differences between Palisota and Pollia are sufficiently great that no recent common ancestor may be inferred. This conclusion is supported by morphological and anatomical data, and we cannot agree with Brenan ( 1966)that these genera should be placed in the same “Group”. Evolutionary direction in aneuploid series Aneuploidy is frequent in the Commelinaceae, and several genera have multiple basic numbers. Aneuploid series of four or more basic numbers occur in Aneilema (cf. above), Commelina, Cyanotis, Cymbispatha and Murdannia. With the exception of Cymbispatha, hypotheses concerning the evolutionary direction within these series have been based on the frequency of the basic numbers. The most frequent basic number has been assumed to have given rise to the less common ones (Lewis, 1964; Morton, 1956; Rao et al., 1972).We can see no inherent reason why this should be true. We find that in Aneilemu, the most primitive species have the basic numberx = 16, and that this has generated a descending series of basic numbers. The most frequent basic number is x = 13, which Rao et al. (1972)incorrectly assumed had given rise to the others. Correlated with the decrease in basic number is a general increase in chromosome size (Faden, 1975). The pattern of basic numbers in Cyanotis, for which there is no recent monograph, appears to be similar to that in Aneilema. The most symmetric karyotypes and smallest chromosomes are found among the species with the uncommon basic number x = 13, not among those with x = 12, the most frequent basic number in the genus (cf.Jones & Kukkonen, 197 1 : fig. 2; and Rao et al., 1972, for a summary of chromosome numbers in Cyanotis). The occurrence, in Belosynapsis, (which is probably the direct ancestor of Cyanotis) of yet more symmetric karyotypes and still smaller chromosomes than in any Cyanotis species (cf.Jones &Jopling, 1972: fig. 3D;Jones and Kukkonen, 197 1 : fig. 3) suggests that symmetric karyotypes and small chromosomes as well as the basic number x = 13 are primitive in Cyanotis. Significantly, these Belosynapsis species also have the basic number x= 13. At the other extreme, the most asymmetric karyotypes and largest chromosomes in Cyanotis are present in the two species with the basic number x = lO(Jones&Jopling, 1972: fig. 3F). The direction of aneuploid change in basic number in Murdannia has probably also been towards reduction, but the evidence is fragmentary. In one evolutionary line, M . gigantea (Vahl) Bruckn., with the basic number x = 11, is apparently less advanced on morphological grounds than the closely related M . simplex, M . lorformis (Hassk.) Rao & Kammathy and M . nud9ora (L.) Brenan, all based on x = 10 (cf. Rao et al., 1972, for a summary of chromosome numbers in Murdannia), suggesting a reduction in basic number. In other lineage, M . edulis (Stokes)Faden (synonym M . scapiJora (Roxb.) Royle), with the reported chromosome number n = 9 (and presumably the basic number x = 91, is morphologically more specialized than such seemingly related taxa as M . divergens (C. B. Clarke) Bruckn., M . hookeri (C. B. Clarke) Bruckn. and M . zeylanica, all of which are based on x = 10 (Faden, 1980).Again, this would imply an aneuploid reduction in basic number. The two closely related species of Murdannia with the basic number x = 6 have several specialized characters such as fused stamen and staminode filament bases, which set them apart fi-om all other species in the genus. Their unusual zyxwv zyxwv zyxwvutsrq zyxwv zyxwv CYTOTAXONOMY O F COMMELINACEAE 32 1 morphological features and mainly African distributions suggest that they are derived from some more generalized Murdannia stock. They most closely resemble the Indian M. koenigii (C. B. Clarke) Bruckn., the chromosomes ofwhich have not been counted. Too little is known about interspecific relationships in Commelina to make any definite assertions about evolutionary direction in the aneuploid series of basic numbers. N o two species with different basic numbers have been shown to be closely related to one another, so it has not been possible to correlate change in basic number with other characters. Even the two basic numbers reported in C. dfluusa (Table 1) have not been investigated in detail. The case of Cymbispatha requires special attention. In this genus there are basic numbers of x = 6, 7 , 8 , 11 and 15 (Jones, 1977, 1978). Jones (1977, 1978) has pointed o u t that somatic chromosome numbers are not useful for the elucidation of the evolutionary pattern in this genus. He has shown that in Cymbispatha the total number of major chromosome arms in the complement ‘nombre fondamental’ is a better indicator of ploidy; it is always a multiple of seven. Karyotypic changes in the genus have occurred by means of Robertsonian fusion of acrocentric or telocentric chromosomes to form metacentrics, a process which has been accompanied by chromosome doubling at various points (Jones, 19 7 7, 197 8). Thus, Robertsonian changes in Cymbispatha always have resulted in a reduced chromosome number and increased chromosome size, a pattern not unlike that described above for Aneilema and Cyanotis, genera wherein the mechanisms of chromosome change have not been investigated. However, a major difference is that aneuploid reductions in Cymbispatha have occurred mainly at polyploid levels, whereas in the other two genera, as well as in Commelina and Murdannia, such changes have happened almost exclusively at the diploid level. I t is because of this difference, as well as the mechanical way in which basic numbers are calculated, that somatic chromosome numbers can be used to assess evolutionary relationships in the latter group of genera but not in Cymbispatha. Finally, it may be noted that karyotype evolution in Cymbispatha has proceeded in the direction of increasing karyotypic symmetry. In Aneilema and Cyanotis, in contrast, the most asymmetric karyotypes appear to be the most advanced. I t is evident that no generalizations about evolutionary trends in karyotype symmetry may be made for the Commelinaceae as a whole. zyxwvutsrq zyxwvut zyxwvuts ACKNOWLEDGEMENTS We wish to thank Dr G. Davidse, MrJ. Lavranos, DrJ. M. Lock, Mr M. G. Price and Dr T. Pocs for supplying living plants; D r W. H. Lewis and Washington University, St. Louis, for furnishing research facilities; Dr P. H. Raven and the Missouri Botanical Garden for providing greenhouse space; the Government of Kenya for granting permission for one of us (R. B. F.) to do research in that country. We are grateful to Drs W. C. Burger, K. Jones and P . Goldblatt for reading portions of the manuscript. Field work in Africa by one of us (R. B. F.) was supported by a National Science Foundation Grant for Improving Doctoral Dissertations in the Field Sciences (GB-40817). The Ministry of Education ofJapan provided full financial support for one of us (Y. S . ) to stay at Washington University, 197 5-76. zyxw 322 zyxwvutsrqponm zyxw zyxwvutsrq R. B. FADEN AND Y. SUDA REFERENCES zyxwv ANDERSON, E. & SAX, K., 1936. A cytological monograph of the American species of Tradescantia. Botanical Garrtte, 9 7: 4 3 3 4 7 6 . BHATTACHARYA, B., 1975. Cytological studies of some Indian members of Commelinaceae. Cytologia, 40: 285-299. BRENAN, J. P. M., 1952. Notes o n African Commehaceae. Kew Bulletin, 7: 179-208. BRENAN, J . P. M., 1966. The classification of Commelinaceae.Journal ofthe Linnean Society, Botany, 59: 349-370. CHIMPHAMBA, B. B., 1973. Chromosome numbers in Malawian Commehaceae. Botanical Journal of the Linnean Society, 6 6 : 303-306. DARLINGTON, C. D., 1929a. Chromosome behaviour and structural hybridity in the Tradescantiae. Journal of Genetics, 21: 207-286. DARLINGTON, C. D., 1929h. Ring-formation in Oenothera and other genera. J o u d @Genetics, 20: 345-363. DARLINGTON, C. D., 1932. Recent Advances in Cytology. London: Churchill. DARLINGTON, C. D., 1938. Chromosome behaviour and structural hyhridity in the Tradescantiae, 11.Journal of Genetics, 35: 259-280. FADEN, R. B., 1974. Commelinaceae. In A. D. Q. Agnew, Upland Kenya Wild Flowers: 653-668. London: Oxford University Press. FADEN, R. B., 1975. A Biosystematic Study ofthe Genus Aneilema R . Br. (Commelinuceae). Ph.D. Thesis, Washington University, St Louis. FADEN, R. B., 1980. The taxonomy and nomenclature of some Asiatic species of Murdannia (Comnielinaceae): The identity of Commelina medica Lour. and Commelina tuberosa Lour. Twon, 29: 7 1-83. FUJISHIMA, H., 1969. Karyological studies in Commelinaceae, I. Karyotypes of Commelina benghalensis L. and Zebrinapurpusii. La Kromosomo, 76: 2486-2493. FUJISHIMA, H., 1970a. Karyological studies in Commelinaceae 11. Karyotypes of Polliajaponica and Aneilema Keisak. Botanical Magazine, Tokyo, 83: 21-26. FUJISHIMA, H., 1970b. Some new findings o n the chromosomes of Polka japonica Thumb. [sic] and Aneilema Keisak Hassk. Chromosome Informotion Service, 1 1 : 6-8. FUKUMOTO, K., 1964. Examination of the chromosome numbers and karyotypes of some Commelina species and formae. Journal ofJapanese Genetics, 39: 340. GANGULY, J . K., 1946. The somatic and meiotic chromosomes of Commelina benghalemis Linn. Current Science, zyxwvu I S : 112. GOLDBLATT, P., 1980. Polyploidy in angiosperms: Monocotyledons. In W. H. Lewis (Ed.), Polyploidy: 219-240. New York: Plenum Press. GUERVIN, C. & LE COQ, C., 1966. Caryologie des Commklinacees: Application a quelques problPnies relatifs a leur wolution. Revue de Cytologie et de Bwlogie VFgitales, 29: 267-328. HARVEY, M. J., 1966. In: IOPB chromosome number reportsVI1. Taxon, 15: 155-163. HEITZ, B., 1968. Commelinacees. 11. Infoormatiom Annuelles de Caryosystematiqueset Cytoghetiques, 2: 26-32. HEITZ, B., 1969. Commelinacees. Informatiom Annuelles de Caryosystematiqueset Cytoghetiques, 3 : 9-12. HSU, C., 1967. Preliminary chromosome studies o n the vascular plants ofTaiwan (I). Taiwanra, 13: 117-129. HSU, C., 197 1. Preliminary chromosome studies o n the vascular plants ofTaiwan (IV): Counts and systematic notes o n some monocotyledons. Taiwania, 16: 123-136. HSU, C., 1972. Preliminary chromosome studies o n the vascular plants of Taiwan (V): Cytotaxononiy o n some monocotyledons. Taiwania, 17: 48-65. HUNT, D. R., 197 1. Dichorisandrathyrstflora.Curtis's Botanical Magazine, n.s., 178:Tab. 590. HUNT, D. R., 1975. The reunion of Setcreasea and Separotheca with Tradescantia. American Commelinaceae: I . Kew Bulletin, 30: 443-458. JONES, K., 1977. The role of Robertsonian change in karyotype evolution in higher plants. In A. de la Chapelle & M. Sorsa (Eds), Chromosomes Today, 6 : 12 1-129. Amsterdam: EIseviedNorth Holland. JONES, K., 1978. Aspects of chromosome evolution in higher plants. In H. W. Woolhousr (Ed.), Advances in Botanical Research, 6 : 119-194. London, New York, San Francisco: Academic Press. JONES, K . & JOPLING, C., 1972. Chromosomes and the classification of the Commelinaceae. BotanicalJournal ofthe Linnean Society, 6 5 : 129-162. JONES, K. & KUKKONEN, I., 1971. The comparative cytology of some Cyanotis species. J o u m l ofthe Indian Botanical Society, 50A: 332-339. KAMMATHY, R. V. & RAO, R. S., 1961a. Notes o n Indian Commelinaceae-11: Cytological observation. Bulletin of the BotanicalSurvey OfIndia, 3 : 167-169. KAMMATHY, R. V. & RAO, R. S., 1961h. Notes o n Indian Commelinaceae-111: Cytological obsrrvations. Bulletin ofthe Botanical Survey of India, 3 : 393-394. UMMATHY, R. V. & RAO, R. S . , 1964. Notes o n Indian Commelinaceae-IV: Cyto-taxonomic observations. Bulletin ofthe Botantcol Survey ofIndia, 6 : 1-6. KAMMATHY, R. V. & RAO, R. S., 1965. In: IOPB chromosome number reports V. Taxon, 14: 191-196. LEWIS, W. H., 1964. Meiotic chromosomes in African Commelinaceae. Sida, I : 274-293. LEWIS, W. H., STRIPLING, H. L. &ROSS, R. G., 1962. Chromosome numbers for some angiosperms of the southern United States and Mexico. Rhodora, 6 4 : 147-161. LEWIS, W. H., SUDA, Y. & OLIVER, R. L., 1967. In: Chromosome numbers of phanerogams. 2. Annals of the Missouri Botanical Garden, 5 4 : 1 7 8- 18 1. zyxwvutsrq zyx zyxwvutsrqp zyxw zyxw zy zyx zyxwvutsr zyxwvut zyxwvuts CYTOTAXONOMY OF COMMELINACEAE 323 LEWIS, W. H . & TADESSE, E., 1964. Chromosome numbers in Ethiopian Commelinaceae. Kirkia, -1: 215-215. McCLlNTOCK, B., 1929. A method for making a c e t o - c a m i n smears permanent. Stain TechnoloQ, 1: 53-56. MALlK, C. P., 1961. Chromosome number in some Indian angiosperms: Monocotylrdons. Science and Culture, 27: 197-198. MANGENOT, S. & MANGENOT, G., 1958. Deuxieme liste d e nombres chromosomiqurs nouveaux chez diverses dicotyledones et monocotyledones d'Afrique Occidentale. Bulletin d u Jardin Botanigue de l ' h t , Bruxrlle,, 28: 3 15-349. MANGENOT, S. & MANGENOT, G., 1962. EnquPte sur les nombres chromosomiques d a m u n r collection d'especes tropitales. Revue de Cytologic et de Biologie Vkgttales, 25: 4 11-447. MATSUURA, H . & S U T 6 , T., 1935. Contributions to the idiogram study in phanerogamous plants. I . Journal d t h r Farulty oj.Science, Hokkaido University, Senes 5 , Botany, 5 : 33. MIEGE, J . , 1960. Nornbres chrornosomiques d e plantes d'Afrique Occidentale. Revue de Cytologic et Biologic L'6g6talei, 21: 373-384. MITRA, K. & DATTA, N., 1967. In: IOPB chromosome number reports XlII. Taxon, 16.445-461. MITSUKLJRI, Y . , 1947. Cytological studies in Commelinaceae. I. Chromosomr numbrrr o f Japanese Coiiimelinacea~~.JapanerrJounal ofcenetics, 22: 18- 19. MORTON, J . K., 1956. Cytotaxonomic studies o n the Gold Coast species of the genus Commelzna Linn.Journa1 of thr Linnran Society, Botany, 5 5 : 507-53 1. MORTON, J . K., 1966. A revision of the genus Aneilema R. Brown (Commelinaceae) with a r p t a x o n o n i i c account o f t h e West African species. Journal ofthe Linnean Society, Botany, 59: 43 1-478. MORTON, J. K., 1967. The Comrnelinaceae of West Africa: A biosystematic survey. Journal ofthe Linnean Society, Botany, 6 0 : 167-221. PANIGRAHI, G. & KAMMATHY, R. V., 1961. Studies in the cytomorphologr o f A m i l e m a rensu latu in eastrrn India. Proceedings of the 481h Indian Science Congress (Roorkee), Part 4 : 13-14. PANIGRAHI, G. & KAMMATHY, R. V., 1962. Cytotaxonomic studies in certain species of Commrlina Linn. in castern India. Proceedings ofthe 4Yth Indian Science Congress Association, Calcutta, Part 3 , Abstractr: 329-330. PANIGRAHI, G. & KAMMATHY, R. V., 1963. Studies in the taxonomy a n d cytology of crrtain species of. Anrzlrrna irniu Into in eastern India. Proceedings ofthe Nafional Academy ofSriences, India, Section B , 3 3 : 491-506. PANIGRAHI, G. & KAMMATHY, R. V., 1964. Cytotaxonomic studies in certain species of Cornmeha Linn. in cdstern Indid.Journal d t h e Indian Botanical Society, 43: 2 9 6 3 10. RAGHAVAN, R. S. & RAO, R. S., 1961. Cytological observations o n the Indian species of Comnielinacear. Current Srirnce, 30: 310-3 1 1 RAGHAVAN, R. S. & RAO, R. S., 1965. Notes o n lndian Commelinaceae VI-Cytological obsrrvations. Nudrub /Calcutta). 8: 39-44. RAO, P. N., 1970. Meiotic studies in Cyanotis viliosa Schult. F. Nucleus (Calcutta), 13' 106-110. RAO, R. S. & KAMMATHY, R. V., 1966. Notes o n Indian Commelinaceae. V. Journal o f t h e Linnean Socifty, Bolafty, 5Y: 305-308. RAO, R. S., RAGHAVAN, R. S. & KAMMATHY, R. V., 1972. Biosystematic s t u d i e ~ o n Indian Comnielinacear-The chromosome pattern a n d evolutiona9 trends. Bulletin ofthe Botanical Survey of India, 12: 242-254. ROHWEDER, 0.. 1956. Die Farinosae in der Vegetation von El Salvador. Abhandlungen aur dem Gebiet der Auslandihunde, ihioersitat Hamburg, 61, Reihe C, iVatunuissenschaftn, 18: 1-197. SHARMA, A. K., 1955. Cytology of some of the members of Commelinaceae and its bearing on thr intrrprrtation ofphvlogeny. Genetica, 27: 323-363. SHARMA, A. K. & SHARMA, A,, 1958. Further investigations o n cytology of members of Comnirlinacear with special referenc-e to the role of polyploidy a n d the origin of ecotypes.Joumal ofceneticr, $6:63-84. SHETTY, B. V. & SUBRAMANYAM, K., 1961. Cytology of some taxa of Commelinaceae. Proceedtngr ofthe 48th Indian Science Congretr fRoorkee), Part 3 , Abstracts: 299. SHE?TY, B. V. & SUBRAMANYAM, K., 1962. Cytological studies in Commelinaceae. Nacku, (Cakutta), 5 : 39-50, SIMMONDS, N. W., 1954. Chromosome behaviour in some tropical plants. Heredity, 8 . 139-146. SUNDARARAGHAVAN, R. & RAO, R. S . , 1965. In: IOPB chromosome number reports V. Taxon, 1 4 : 19 1-196. TOMLINSON, P B., 1966. Anatomical data in the classification of Commelinacear. Journal uf the Linnean Society, Botany, i 9 : 37 1-39.5. ZAMAN, M. A. & AHMED, M., 1972. Cytogenetics of Commelinaceae. 1. Meiotic brhaviour and Bchi-oinosornrs in Cumrnelina benghalenszs L. Bangladesh Joumal ofBotany, I : 14 1-148. zyxwv zyxw APPENDIX Collection data and vouchersfor the species counted Anthericopsis sepalosa ( C . B. Clarke) Engl. KENYA. TAITA: 18 k m on Voi-Taveta road from Nairobi-Mombasa road, Faden 6.Faden 741504 ( M O ) . 324 zyxwvu zyxwvut zyxwvu R. B. FADEN AND Y. SUDA Coleotrype bruecknerana Mildbr . Commelina benghalensis L. Commelina bracteosa Hassk. Commelina congesta C. B. Clarke Commelina dgusa Burm. fil. Commelinafoliacea Chiov. subsp. foliacea Commelinafoliacea Chiov. subsp. ‘A’ Commelina lattfolia A. Rich. zy zy zyxwv zyxwvut zyxw Commelina macrospermaJ. K. Morton Commelina thomasii Hutch. Commelina trilobosperma K. Schum. Commelina zambesica C. B. Clarke Commelina sp . ‘E’ of Faden ( 19 74) Cyanotisfoecunda Hassk. Cyanotis Tlyctitropa Deflers CyanotiJ paludosa Brenan Cyanoti., sp. ‘A’ TANZXNIA. M O R O G O R O : Uluguru Mts., Kimboza Forest Reserve, Faden 6. Faden 74/415 (MO). PHILIPPINES. Laguna, College campus, Price2436 (US). KENYA. KILIFI : Mariakani-Kaloleni road, k m 2.6, Faden 6. Faden 71/838 (MO). GHANA. EASTERN : Nyanao Hill near Nsawam, Faden, Faden 6. Lock 74/70 (MO). PHILIPPINES. Laguna, College campus, Price2435 (US). UGANDA. KIGEZI :4-7 mi. u p KaiziBitereko road, Faden, Evans, Lye 6. Lock 69/1114 (EA). TANZANIA. M O R O G O R O : Mkungwe Hill, Uluguru Mts., Faden, Evans 6. Pbcs 70/371 (EA). KENYA. MACHAKOS : Thika-Garissa road, near Kangonde, Faden 6. Faden 74/733 (MO). GHANA. EASTERN: Nyanao Hill near Nsawam, Faden, Faden 6. Lock 74/69 (MO). GHANA. EASTERN : 3 k m SW of Adaiso, Faden, Faden 6. Lock 74/5 5 ( M 01. KENYA. KISUMU: 2.3 kmbeyond Paponditi on the Nyakwere road, Evans 6.Maikweki 13 (MO). KENYA. KWALE: Shimba Hills, Sheldrick’s Falls area, Faden 6.Evans 701436 (MO). KENYA. KAKAMEGA: Kakamega Forest, S side ofYala River, Faden et al. 70120 (EA). KENYA. SOUTH NYANZA: Lambwe Valley, Riamkanga, Faden 69/1304 ( M O ) . SOUTHERN YEMEN. Near summit of Jabal JihZf, Lavranos 6. Bavazzano 8620 (US). KENYA. TRANS NZOIA: Elgon South road, first swamp beyond aerodrome, Faden, Evans 6. Tweedie 69/7 16 ( EA). KENYA. TANA RIVER: MalindiGarsen road, 0.8 k m towards Garsen from turnoff to Kibusu, Faden 6. Faden 7411 I74 (MO). zyx zyx zyx zyxwvutsr zyx zyxwvuts CYTOTAXONOMY OF COMMELINACEAE filoJcopa afrzcana (P. Beduv.) C. B. Cldrkc. subsp.petro/hla J . K. Morton Murdunnm clarkeana Brennn Murdanriza iimplex iVahl) Brenan Murdanrzia zeylarizca (C. B. Clarke) B riic k n . Palisota barteri Hookel Palisota hirsuta (Thunb.) K. Schum. Paliwta onentalz, K. Schuni. Polliu conden,atn C. B. Clarke Pollia secundifEora (Bl.) Bakh. fil. t’ollia thyrs$orn (B1.1 Steud. Poly cpatha paniculata Bcnth. Starlfieldzella irnl/erforata (C. B. Clarke) B reii~iivar. imperforata Stanfieldiella olzgantha ( Mildbr.) Brcnan 22 325 UGANDA. BUNYORO : 23 mi. N of Butiaba turnoff on Butiaba-Murchison Falls road, Faden, Evans 6.Lye 69/1067 (EA). UGANDA. EAST M E N G O : Ssezzibwa Falls, Faden 6.Evans 69/969 (EA). KENYA. NYERI : State Lodge Road, 1 mi. pastjunction with Upper Thego fishingroad, Faden 6.Evans 69/304 (EA). GHANA. CENTRAL: C. 1 km E of WinnebaJunction, Faden, Faden 6.Lock 74/64 ( M O ) . SRI LANKA. KANDY: KandyMahiyanganaroad, mi. 35, 14 N o \ . 1974, Dav2dses.n. ( M O ) . GHANA. EASTERN : Nyanao Hill near Nsawam, Faden, Faden 6.Lock 74/75 (MO). GHANA. EASTERN: Nyanao Hill near Nsawam, Faden, Faden 6.Lock 74/66 (MO). TANZANIA. M O R O G O R O : Uluguru Mts., near Kinole sawmill, P6cs 6874/A (MO). GHANA. EASTERN : Atewa Range Forest Reserve, Sagyimase Track, F a d m etal. 7 4 / 3 7 ( M O ) . TANZANIA. TANGA: Eastern Usambara Mts., crossing of Kwamkuyu River on Muheza-Amani road, Faden’& Faden 74/3 72 ( M 0) . PHILIPPINES. Laguna, College hortorium, spontaneous, Price3126 (US). PHILIPPINES. Laguna, Mt. Makiling, Przce3131 (US) GHANA. EASTERN : Atewa Rangc Forest Reserve, Sagyimase Track, Faden et al. 74/35 ( M O ) . GHANA. EASTERN : Atewa Rangc. Forest Reserve, Sagyimase Track, Faden et al. 74/34 ( M O ) . GHANA. EASTERN : Nyanao Hill near Nsawam, Faden, Faden 6.Lock 74/73 (MO).