American Journal of Botany 85(5): 681±687. 1998.

GENOME SIZE AND KARYOTYPE EVOLUTION IN THE SLIPPER ORCHIDS (:)1

ANTONY V. C OX,2 GREGORY J. ABDELNOUR,MICHAEL D. BENNETT, AND ILIA J. LEITCH

Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey, TW9 3DS, UK

Nuclear DNA contents (4C) were estimated by Feulgen microdensitometry in 27 species of slipper orchids. These data and recent information concerning the molecular systematics of Cypripedioideae allow an interesting re-evaluation of karyo- type and genome size variation among slipper orchids in a phylogenetic context. DNA amounts differed 5.7-fold, from 24.4 pg in longifolium to 138.1 pg in wardii. The most derived clades of the conduplicate-leaved slipper orchids have undergone a radical process of genome fragmentation that is most parsimoniously explained by Rob- ertsonian changes involving centric ®ssion. This process seems to have occurred independently of genome size variation. However, it may re¯ect environmental or selective pressures favoring higher numbers of linkage groups in the karyotype.

Key words: Cypripedioideae; centric ®ssion; genome size; karyotype evolution; Orchidaceae; phylogeny.

Slipper orchids (Cypripedioideae: Orchidaceae) are of genome size data prevented Cox et al. from determin- probably the best characterized of the orchid subfamilies. ing whether the change in chromosome number was ac- Great horticultural importance has contributed signi®- companied by a change in genome size. cantly to prolonged scienti®c interest, not least because To investigate the interrelationship among genome extensive collections have made much material available size, karyotype evolution, and phylogenetic af®nities we for examination. Indeed, studies spanning over 150 years measured nuclear genome sizes for 27 slipper orchid spe- provide a substantial body of information concerning cies sampled from the three largest genera, Phragmipe- their cytology (Karasawa, 1979, 1980, 1986; Karasawa dium, Paphiopedilum, and . The data of Cox and Tanaka, 1980, 1981; Karasawa and Saito, 1982; At- et al. (1997) were used to place genome size and karyo- wood, 1984; Karasawa and Aoyama, 1986, 1988), mor- type data in a phylogenetic context and to suggest mech- phology (Rosso, 1966; Atwood, 1984), (Rei- anisms and directionality of change. chenbach, 1854; Rolfe, 1896; Atwood, 1984; Albert and Pettersson, 1994), molecular systematics (Albert and MATERIALS AND METHODS Chase, 1992; Cox et al., 1997), and pollination biology (Dodson, 1966; Vogt, 1990; BaÈnziger, 1994, 1996; Chris- materialÐRoot tips for chromosome counts and DNA mea- surements were taken from the orchid collection at the Royal Botanic tensen, 1994). Gardens, Kew. Plant accession numbers, vouchers, and karyotypic com- Detailed taxonomic studies have been made of the positions are presented in Table 1. The recent molecular phylogeny of three major genera, Cypripedium (Cox, 1995; Cribb, in Cox et al. (in press) was used to assist sampling. All sections in press), Paphiopedilum (Cribb, 1987), and Phragmipe- Phragmipedium and Paphiopedilum (except section Pardalopetalum) dium (Garay, 1979; McCook, 1989). The two remaining were represented in the study, but reduced sampling in Cypripedium genera, and monotypic , are occurred due to lack of living material. Living material of Mexipedium poorly known, owing to their rarity in nature and, in the and Selenipedium was unavailable for study. case of the former, great dif®culty in maintaining in cultivation. Chromosome countsÐActively growing root tips were collected and Extensive chromosome data are available for slipper pretreated with 0.002 mol/L 8-hydroxyquinoline for 4±5 h at 18ЊC (Kar- orchids. Chromosome counts are published for almost all asawa, 1979) prior to ®xing in freshly prepared 3:1 (v/v) absolute eth- species of Paphiopedilum and Phragmipedium and many anol/glacial acetic acid at 4ЊC for a minimum of 1 h. Root tips were of Cypripedium; detailed karyotypes are available for transferred to 45% acetic acid for 5 min at 4ЊC, hydrolyzed in 1 mol/ most species of the former two genera. Recently, parsi- L HCl at 60ЊC for 15 s (Karasawa, 1979), and stained in the dark with mony analysis of nuclear ribosomal internal transcribed Feulgen solution for a minimum of 30 min and a maximum of 24 h at spacer (ITS) sequence data from nearly 100 slipper or- 4ЊC. The preparation was then squashed under a coverslip in 2% aceto- chids elucidated the phylogeny of slipper orchids (Cox orcein for cytological analysis. et al., 1997). A survey of the karyotype data in the con- DNA measurementsÐThree root tips of each species of slipper or- text of the phylogenetic tree revealed a general trend to- chid and the calibration standard Allium cepa cv. ``Ailsa Craig'' (4C ϭ ward increased chromosome number. However, paucity 67.1 pg; Bennett and Smith, 1976) were collected on the same day and ®xed in freshly prepared 3:1 (v/v) absolute ethanol/glacial acetic acid for a minimum of 24 h at 4ЊC before hydrolysis in 5 mol/L HCl at 25ЊC 1 Manuscript received 28 May 1997; revision accepted 12 September 1997. for 40 min. Root tips were then rinsed in distilled water, stained in The authors thank Lynda Hanson for valuable technical assistance pararosanoline solution for 2 h at 24ЊC, washed in three aliquots of and Mark Chase and Keith Jones for helpful discussions. Sandra Bell SO2-water for 10 min each, transferred to distilled water, and stored in kindly provided root tips of slipper orchids. the dark at 4ЊC for up to 24 h. Finally, the roots were squashed in 45% 2 Author for correspondence ([email protected]). acetic acid. Readings of ten prophase (4C) nuclei on each of three slides 681 682 AMERICAN JOURNAL OF BOTANY [Vol. 85

TABLE 1. Plant material used in this study: accession number, chromosome number, and genome size.a

Cytology reference Chromosome Mean 4C DNA Taxon Accession number number numberb amount (pg) Paphiopedilum Sect. Barbata P. appletonianum (Gower) Rolfe 1981-1534 96-61 38(24) 129.7 P. wardii Summerh. 1990-266 96-66 41(29) 138.1 P. callosum (Rchb.f.) Stein 1990-3253 96-155 32(12) 96.2 P. purpuratum (Lindl.) Stein 1988-2124 96-158 40(28) 108.5 P. barbatum (Lindl.) P®tzer 1990-252 96-62 38(24) 135.0 P. tonsum (Rchb.f.) P®tzer 1988-3165 96-67 32(12) 112.6 M. mastersianum (Rchb.f) Stein 1993-522 96-63 36(20) 118.9 P. sukhukulii Schoser & Senghas 1990-265 96-65 40(28) 118.9 P. urbanianum Fowlie TP4 2653 96-87 40 Ð Sect. Coryopedilum P. philippinense (Rchb.f.) Stein 1990-198 96-107 26 93.0 P. rothschildianum (Rchb.f.) Stein 1986-2479 96-108 26 90.3 P. glanduliferum (Blume) Stein 1953-38501 96-159 26 94.9 P. dianthum T. Tang & F. T. Wang 23-10-90 GREL 96-85 26 Ð Sect. Paphiopedilum P. insigne (Wall. ex Lindl.) P®tzer Ð Ð 26 92.1c P. druryi (Bedd.) Stein 1976-952 96-74 30(8) 106.5 P. gratrixianum (Masters) Guill. 1979-0975 96-160 26d 100.0 P. villosum (Lindl.) Stein Ð Ð 26 89.9e Sect. Cochlopetalum P. primulinum M. Wood & Taylor 1981-1628 95-156 32(14) 83.6 P. victoria-mariae (Rolfe) Rolfe 1988-3173 96-84 36 85.6 Sect. Brachypetalum P. concolor (Lindl.) P®tzer 1988-1641 96-76 26 77.9 P. godefroyae (Godefr.-Lebeuf) Stein 1988-1642 97-77 26 71.2 Sect. Parvisepalum P. micranthum T. Tang & F. T. Wang 1990-195 96-73 26 91.0 P. delenatti Guill. 1989-3408 96-72 26 87.3 Phragmipedium Sect. Phragmipedium P. caudatum (Lindl.) Rolfe 1986-2307 96-69 28(20) 36.7 Sect. Platypetalum P. lindleyanum (Lindl.) Rolfe 1983-341 96-71 22(10) 32.1 Sect. Lorifolia P. pearcei (Rchb.f.) Rauh & Senghas 1987-4033 96-154 20,21,22 25.3 P. longifolium (Rchb.f. & Wasc.) Rolfe 1976-2035 96-70 21(5) 24.4 Sect. Micropetalum P. besseae Dodson & J. Kuhn 1992-3664 96-83 24 28.3 Cypripedium C. pubescens Willd. 1991-1118 96-103 20 129.5 C. formosanum Hay 1986-684 96-105 20 113.9 C. californicum A. Gray K541 96-106 N/A 86.2 a Dash indicates data not available. b Number of telocentric chromosomes (where known) shown in parentheses. c Cox et al. (1993). d New chromosome record. e Narayan, Parida, and Vij (1989).

were made using a Vickers M85 scanning microdensitometer (Vickers, Genome sizeÐThe genome sizes of 27 slipper orchids, UK) and mean values calculated. measured by Feulgen microdensitometry, ranged from 24.4 pg in Phragmipedium longifolium to 138.1 pg in RESULTS (Table 1). Table 1 also includes the genome sizes of and P. vil- Chromosome numberÐA chromosome count was losum previously reported by Cox et al. (1993) and Na- made for one species of Paphiopedilum not previously rayan, Parida, and Vij (1989), respectively. The 4C DNA determined (P. gratrixianum) and is listed in Table 1. amounts for three Cypripedium species ranged from 86.2 May 1998] COX ET AL.ÐKARYOTYPE EVOLUTION IN SLIPPER ORCHIDS 683

TABLE 2. Range in chromosome number, karyotype structure (where known), nombre fondamental (n.f.; Mathey, 1949), and range of 4C DNA amount. (Chromosome data taken from Karasawa [1979]).

Range in 4C DNA Genus Section 2n n.f. amount (pg) Paphiopedilum Barbata 28 to 42 52±54 96.2±138.1 Pardalopetalum 26 52 Ð Coryopedilum 26 52 90.3±94.9 Paphiopedilum 26 (excluding P. druryi 52 89.9±106.5 and P. spicerianum) Cochlopetalum 32±37 50 83.6±85.6 Brachypetalum 26 52 71.2±77.9 Parvisepalum 26 52 87.3±91.0 Phragmipedium Phragmipedium 28 36 36.7 Platypetalum 22 34 32.1 Lorifolia 18±21 36±37 24.4±25.3 Micropetalum 24, 30 36 28.3 Cypripedium Ð 20 40 86.2±129.5 to 129.5 pg. In congruent molecular phylogenies using for Cypripedium and the subgeneric groupings for Pa- rbcL (Albert and Chase, 1992) and ITS sequence data phiopedilum (sensu Cribb, 1987) and Phragmipedium (Cox et al., 1997) Cypripedium was sister to the two con- (McCook, 1989). This study adds a further 27 DNA es- duplicate-leaved genera, Paphiopedilum and Phragmi- timates for Cypripedioideae, making it the most studied pedium. Our results show that Paphiopedilum species had orchid subfamily for genome size. genome sizes spanning a similar range to that of Cypri- The present work has considerably increased the range pedium, (i.e., 71.2±138.1 pg), whereas those in Phrag- of genome sizes estimated in Orchidaceae from 36-fold mipedium were much smaller; 4C DNA amounts in (based on previous published data, see above) to 57-fold, Phragmipedium ranged from 24.4 to 36.7 pg. making it the third most variable monocotyledonous fam- Although large genome size variation exists at the fam- ily after Poaceae (103.9-fold) and Iridaceae (66.1-fold). ily level, this is not the case at subfamily and genus levels Considering the species richness of Orchidaceae, it seems (Tables 1, 2). In the 27 species of Cypripedioideae mea- likely that greater sampling of genera for genome size sured, 4C DNA amounts differed 5.7-fold. At the generic estimates will increase this range further. level, all three genera examined showed less than twofold variation in genome size (Paphiopedilum 1.9-fold, Karyotype structure and genome size evolution in Cy- Phragmipedium 1.5-fold, and Cypripedium 1.5-fold), pripedioideaeÐThe comprehensive ITS rDNA molecular which is low compared with many other genera examined phylogeny for the slipper orchids (Cox et al., 1997) has in the angiosperms (M. D. Bennett and I. J. Leitch, RBG provided clari®cation to the sometimes contentious sub- Kew, personal communication). Within the genus Pa- familial taxonomy of Cypripedioideae. This molecular phiopedilum, the most variable sections were Barbata phylogeny, in conjunction with previously published (1.4-fold) and Paphiopedilum (1.2-fold). The remaining karyotypic data and the new genome size data reported sections showed minimal variation. Given the wide sam- here, provides important insights into the molecular as- pling of Paphiopedilum species from all but section Par- pects of genome evolution in the slipper orchids because dalopetalum, this genus seems unlikely to show consid- it enables trends in karyotype structure and genome size erable variation in genome size. to be interpreted in a phylogenetic context. In view of the evolutionary signi®cance of such studies, it is sur- DISCUSSION prising that more analyses of this type have not been Genome size variation in the OrchidaceaeÐAlthough reported; only one other similar investigation (Sessions genome size data are available for ϳ1% of the angio- and Larson, 1987), detailing genome evolution in sala- sperm ¯ora (Bennett and Leitch, 1995), information for manders, is known to us. Orchidaceae, arguably the largest angiosperm family with Cox et al. (1997) identi®ed a correlation between the an estimated 19 000 species (Dressler, 1993) is underre- degree of evolutionary ``advancement'' and an increase presented (Bennett and Leitch, 1995). Previously, ge- in chromosome number across the slipper orchids as a nome size data were available for 46 orchid species (Ca- whole. However, the new data presented here reveal that pesius et al., 1975; Capesius, 1976; Nagl and Capesius, within each of the major genera (Cypripedium, Phrag- 1977; Narayan, Parida, and Vij, 1989; Arumuganathan mipedium, and Paphiopedilum) the mode of genome evo- and Earle, 1991; Cox et al., 1993; V. A. Albert et al., lution may be different. New York Botanic Garden, personal communication). Figure 1 shows the 4C DNA amounts of the studied Karyotype and genome size evolution in Cypripedi- species mapped onto the ITS phylogenetic tree of Cox et umÐThe genus Cypripedium is characterized by a con- al. (1997). This tree is congruent with rbcL results (Al- stant diploid karyotype of 20 metacentric chromosomes. bert, 1994), although sampling is more complete in the Despite this constancy, the genome size in the three taxa ITS study. Table 2 summarizes the karyotype structure, measured showed a 1.5-fold range in DNA amount (4C range of diploid chromosome numbers, and genome sizes ϭ 86.2±129.5 pg). 684 AMERICAN JOURNAL OF BOTANY [Vol. 85

Fig. 1. Variation of 4C DNA amount (diamonds) in Cypripedioideae. Generic boundaries are shown as dotted lines; diploid chromosome numbers for subgeneric groups are shown in boxes. The phylogenetic tree, using Vanilla as the outgroup (not shown), is that reported by Cox et al. (1997), but modi®ed to show only taxa for which relevant DNA amount and/or cytological data are available.

Karyotype evolution in Paphiopedilum and Phrag- mipedium and Paphiopedilum and that karyotype ortho- mipediumÐIn contrast to Cypripedium, wide sampling selection is operating in some species groups. The C- within the major subgeneric, monophyletic sections of banding studies of Karasawa and Tanaka (1980) support Phragmipedium and Paphiopedilum indicate a trend for this hypothesis, as do in situ hybridization studies with chromosome number to increase with evolutionary ad- telomere-speci®c probes, which failed to identify inter- vancement. Our new count for Paphiopedilum gratrix- stitial telomeric repeat sequences that may have remained ianum augments the data of Karasawa (1979, 1980, from fusion events (Cox et al., 1993). 1986), Karasawa and Aoyama (1986, 1988), Karasawa In Phragmipedium, the diploid chromosome number and Saito (1982), and Karasawa and Tanaka (1980, ranges from 2n ϭ 18 (all metacentric chromosomes, e.g., 1981). Chromosome number information for the genus is P. boissierianum, section Platypetalum)to2n ϭ 30 (six now almost complete. Karyotype analysis also revealed metacentric and 24 telocentric, e.g., P. schlimii, section that, despite a range of chromosome numbers, the total Micropetalum), yet the total number of chromosome arms number of chromosome arms in a karyotype (nombre (i.e., n.f.) of 34±37 is broadly conserved (Karasawa, fondamental or n.f.; Mathey, 1949) was almost complete- 1980). The phylogenetic tree indicates that a diploid ly conserved within genera (Table 2). Such conservation karyotype of 18 metacentric chromosomes is the ple- of n.f. strongly suggests a Robertsonian relationship be- siomorphic condition (retained in P. boissierianum) and tween the different karyotypes whereby changes in chro- that centric ®ssion generated karyotypes with increasing mosome number are generated by the ®ssion or fusion of numbers of telocentric chromosomes. chromosomes at or near the centromere to generate either A similar situation is apparent in the genus Paphio- telocentric or metacentric chromosomes, respectively pedilum. The most common diploid chromosome number (Robertson, 1916). Duncan and MacLeod (1949, 1950) is 26 metacentric chromosomes (n.f. ϭ 52) as displayed were the ®rst to propose Robertsonian change to explain in sections Parvisepalum and Brachypetalum, the ®rst the maintenance of arm number in these two genera. two successive sister groups of the other sections and also From the phylogenetic information of Cox et al. (1997) in sections Pardalopetalum, Coryopedilum, and Paphio- it is most parsimonious to postulate centric ®ssion of pedilum (excluding P. druryi and P. spicerianum), which metacentric chromosomes into telocentrics as the pre- have 30 chromosomes, of which eight are telocentric, re- dominant mechanism of karyotype evolution in Phrag- taining the n.f. of 52; Table 2). Species in sections Coch- May 1998] COX ET AL.ÐKARYOTYPE EVOLUTION IN SLIPPER ORCHIDS 685

Fig. 2. The relationships of diploid chromosome number, 4C DNA amount (diamonds), and number of metacentric chromosomes (squares) in the complement of Paphiopedilum species of known genome size. Lines indicate trends derived from regression analysis. lopetalum and Barbata depart from this stable karyotypic In contrast, Paphiopedilum species not only have the pattern with diploid numbers usually ranging from 2n ϭ largest genomes for the subfamily but also exhibit the 32 to 36 (n.f. ϭ 50) in section Cochlopetalum, and 2n ϭ largest range in 4C amounts (71.2±138.1 pg; 1.9-fold 28 to 42 (n.f. ϭ 52) in section Barbata. As in Phrag- range). Despite the differences both genera show a trend mipedium the phylogenetic information of Cox et al. toward increased genome size with evolutionary ad- (1997) indicates that the plesiomorphic karyotype for Pa- vancement (Fig. 1), although there are still too few data phiopedilum comprised 26 metacentric chromosomes for Phragmipedium species to offer robust support for with increases in chromosome number achieved predom- this correlation. inantly through centric ®ssion. Centric ®ssion seems to The suggested mode of chromosome evolution through have been most common in section Barbata in which centric ®ssion would entail additional DNA being added karyotypes of some species are dominated by telocentric to the genome as the metacentric chromosomes split into chromosomes (e.g., P. sukhakulii with 2n ϭ 40 of which telocentrics (Fig. 2). The need to synthesize telomeric 28 are telocentric, n.f. ϭ 52). The ``ancestral'' karyotype DNA sequences to stabilize the new chromosome ends of section Cochlopetalum appears to have lost either a following ®ssion is one likely source of the additional single metacentric or two telocentric chromosomes prior DNA (Werner, Kota, and Gill, 1992). However, most of to divergence of extant species in this section, which the additional DNA probably consists of many different would account for a n.f. of 50 rather than 52. types of repetitive DNA sequences (e.g., Smyth, 1991). In plant genera showing increases in genome size with Genome size evolution in Paphiopedilum and Phrag- evolutionary advancement (based on nonmolecular mark- mipediumÐPaphiopedilum and Phragmipedium differ ers) this is sometimes achieved nonrandomly either by considerably in genome size. The many morphological addition of equal amounts of DNA to all chromosomes similarities between these two conduplicate-leaved gen- in the complement (e.g., Vicia, Raina, 1990; Lathryus, era, their similar modes of chromosome evolution, and Narayan and Durrant, 1983; Nicotiana, Narayan, 1987) the fact that they inhabit similar tropical latitudes and or by unequal addition of DNA to chromosomes, leading habitats are all in stark contrast to their extreme differ- to a change in karyotype symmetry (e.g., Oxalis, de Az- ences in genome size. All Phragmipedium species have kue and MartõÂnez, 1988; Cypella and Hesperoxiphion, small genomes relative to Cypripedioideae as a whole Kenton et al., 1990). An examination of karyotypes of and exhibit a narrow range in 4C values (24.4±36.7 pg). Paphiopedilum in a phylogenetic context cannot deter- 686 AMERICAN JOURNAL OF BOTANY [Vol. 85 mine unequivocally where the extra DNA is being added presently unknown. In a study of the cycad genus Zamia or indeed which chromosomes are undergoing ®ssion. L., which exhibits extensive centric ®ssion, the onset of However, Atwood (1984) found that differences in arm stressful ecological conditions was suggested as the se- size are signi®cantly larger in telocentrics than in meta- lective pressure that resulted in Robertsonian chromo- centrics in section Barbata with three possible excep- some change (Caputo et al., 1996). In the slipper orchids, tions. This suggests that large chromosomes may function Cox et al. (1997) observed that many of the species with better in cells as telocentrics or that centromeres break high chromosome numbers (i.e., high numbers of telo- too easily in large chromosomes (i.e., they are ``fragile''). centrics) were narrow endemics on Malaysian islands These questions can only be answered once the additional (Cribb, 1987). The rigorously selective environments to DNA has been isolated, characterized, and physically lo- which island ¯oras may be subjected could contribute to calized using molecular approaches such as in situ hy- the selection pressures acting on genomes, causing pref- bridization. erential selection of karyotypes that undergo centric ®s- The only previous study known to the authors involv- sion. A greater number of chromosomes could provide a ing changes in DNA amount with centric ®ssion was re- means to increase recombination and therefore genetic ported by de Azkue and MartõÂnez (1988) in Oxalis. Here, variation, enabling these slipper orchids to adapt more as in Paphiopedilum, centric ®ssion of chromosomes was rapidly to their environment (Stebbins, 1971). accompanied by an increase in DNA content, although A combined approach using molecular sequence data, the nature and location of the extra DNA were not dis- cytological observations, and DNA amount measure- cussed. The two studies that have compared DNA ments can clearly provide a more robust and comprehen- amounts in species presumed to have arisen by Robert- sive insight into the mechanisms and constraints that con- sonian fusion indicate opposite trends in each case. Ni- trol genome evolution than individual studies can. These shikawa, Furuta, and Endo (1979) found that DNA con- areas of research display such rich interplay that we can- tent increased in several Lycoris species as the chromo- not fully understand processes controlling genome (and some number decreased through centric fusion. In con- therefore organismal) evolution by pursuing more narrow trast, Pijnacker and Schotsman (1988) reported that DNA studies. was lost following centric fusion of Callitriche chromo- somes. Together, the available studies indicate that there LITERATURE CITED is no clear trend in genome size change associated with Robertsonian ®ssion/fusion, suggesting that, despite ap- ALBERT, V. A., 1994. Cladistic relationships of the slipper orchids (Cy- parent similarity in karyotype evolution, the genomic pripedioideae: Orchidaceae) from congruent morphological and molecular data. Lindleyana 9: 115±132. mechanisms involved may be different in each case. The , AND M. W. CHASE. 1992. Mexipedium: a new genus of slipper two are most likely independent. orchid (Cypripedioideae: Orchidaceae). Lindleyana 7: 172±176. Although both conduplicate-leaved genera appear to ÐÐ, AND B. PETTERSSON. 1994. Expansion of genus Paphiopedilum to have undergone a similar centric ®ssion process, we can- include all conduplicate-leaved slipper orchids (Cypripedioideae: not yet determine whether these events have occurred Orchidaceae). Lindleyana 9: 133±139. independently. Such co-occurrence may be due to the in- ARUMUGANATHAN, K., AND E. D. EARLE. 1991. Nuclear DNA content of some important plant species. Plant Molecular Biology Reporter heritance of a common, ``fragile'' centromere that origi- 9: 208±218. nated in a common ancestor. If fragmentation is the result ATWOOD, J. T., JR. 1984. The relationships of the slipper orchids (sub- of a fragile site at or near the centromere of ancestral family Cypripedioideae, Orchidaceae). Selbyana 7: 129±147. slipper chromosomes, that fragile site is not evident in BAÈ NZIGER, H. 1994. Studies on the natural pollination of three species Cypripedium. Since fragmentation has occurred in both of wild lady-slipper orchids (Paphiopedilum) in Southeast . In conduplicate-leaved genera, it is unlikely to have been A. Pridgeon [ed.], Proceedings of the 14th World Orchid Congress, 201±202. HMSO, Edinburgh. in¯uenced by changes in genome size; Phragmipedium . 1996. The mesmerizing wart: the pollination strategy of epi- karyotypes are as extensively fragmented as those of Pa- phytic lady slipper orchid (Lindl.) Stein phiopedilum but contain ϳ25% of the DNA. Cytological (Orchidaceae). Botanical Journal of the Linnean Society 121: 59± studies of Mexipedium, the sister taxon to Phragmipe- 90. dium, may help to determine whether centric ®ssion has BENNETT,M.D.,AND I. J. LEITCH. 1995. Nuclear DNA amounts in arisen independently in both lineages. If Mexipedium has angiosperms. Annals of Botany 76: 113±176. , AND J. B. SMITH. 1976. Nuclear DNA amounts in angiosperms. a complement of metacentric chromosomes, then expres- Philosophical Transactions of the Royal Society of London B 274: sion of the fragile site in the two conduplicate-leaved 227±274. genera is most parsimoniously explained as an indepen- CAPESIUS, I., 1976. Isolation and characterization of native AT-rich sat- dently acquired condition. We are currently identifying ellite DNA from nuclei of the orchid Cymbidium. FEBS Letters 68: chromosomal markers that will help to clarify the situa- 255±258. tion. ,B.BIERWEILER,K.BACHMANN,W.RUÈ CKER, AND W. N AGL. 1975. An A ϩ T-rich satellite DNA in a monocotyledonous plant, Overall, the data indicate that although genome size Cymbidium. Biochimica et Biophysica Acta 395: 67±73. varies 5.7-fold in slipper orchids, it does not appear to CAPUTO, P. S., G. L. COZZOLINO,A.MORETTI, AND D. W. STEVENSON. apply constraints to the other mechanisms of karyotype 1996. Karyology and phylogeny of some Mesoamerican species of evolution. This is particularly apparent in the two con- Zamia (Zamiaceae). American Journal of Botany 83: 1513±1520. duplicate-leaved genera, which differ considerably in ge- CHRISTENSEN, D. E. 1994. Fly pollination in the Orchidaceae. In J. Ar- nome size but have undergone similar types of Robert- ditti [ed.], Orchid biology: reviews and perspectives, VI, 415±454. John Wiley and Sons, New York, NY. sonian change to increase the number of linkage groups COX, A. V. 1995. The utility of 5S rDNA in phylogenetic reconstruc- in the karyotype. The selection pressures acting to bring tions. Ph.D. dissertation, University of Reading, Reading. about this particular mode of chromosome evolution are ,S.T.BENNETT,A.S.PAROKONNY,A.KENTON,M.A.CALI- May 1998] COX ET AL.ÐKARYOTYPE EVOLUTION IN SLIPPER ORCHIDS 687

MASSIA, AND M. D. BENNETT. 1993. Comparison of plant telomere MATHEY, T. 1949. Les chromosomes des VerteÂbreÂs. F. Rouge. Lausanne. locations using a PCR-generated synthetic probe. Annals of Botany MCCOOK, L. 1989. Systematics of Phragmipedium (Cypripedioideae: 72: 239±247. Orchidaceae). Ph.D. dissertation, Cornell University, Ithaca, New ,A.M.PRIDGEON,V.A.ALBERT, AND M. W. CHASE. 1997. Phy- York. logenetics of the slipper orchids (Cypripedioideae: Orchidaceae): NAGL,W.,AND I. CAPESIUS. 1977. Repetitive DNA and heterochromatin nuclear rDNA ITS sequences. Plant Systematics and Evolution 208: as factors of karyotype evolution on phylogeny and ontogeny of 197±223. orchids. Chromosomes Today 6: 141±152. CRIBB, P. J. 1987. The genus Paphiopedilum. Collingridge, London. NARAYAN, R. K. J. 1987. Nuclear DNA changes, genome differentiation . In press. The genus Cypripedium. Royal Botanic Gardens, and evolution in Nicotiana (Solanaceae). Plant Systematics and Kew. Evolution 157: 161±180. DE AZKUE, D., AND A. MARTÂõNEZ. 1988. DNA content and chromosome , AND A. DURRANT. 1983. DNA distribution in chromosomes of evolution in the shrubby Oxalis. Genome 30: 52±57. Lathyrus species. Genetica 61: 141±154. DODSON, C. H. 1966. Studies on orchid pollination. Cypripedium, ,A.PARIDA, AND S. P. VIJ. 1989. DNA variation in Orchidaceae. Phragmopedium [sic] and allied genera. American Orchid Society Nucleus 32: 71±75. Bulletin 35: 125±128. NISHIKAWA, K., Y. FURUTA, AND H. ENDO. 1979. Consideration of the DRESSLER, R. 1993. Phylogeny and classi®cation of the orchid family. chromosome evolution on the basis of nuclear DNA content and Cambridge University Press, Cambridge. total chromosome length in Lycoris. Japanese Journal of Genetics DUNCAN, R. E., AND R. A. MACLEOD. 1949. The chromosomes of the 54: 387±396. continental species of Paphiopedilum with solid green leaves. PIJNACKER,L.P,AND H. D. SCHOTSMAN. 1988. Nuclear DNA amounts American Orchid Society Bulletin 18: 84±89. in European Callitriche species (Callitrichaceae). Acta Botanica , AND . 1950. The chromosomes of Eremantha tesselata. Neerlandica 37: 129±135. American Orchid Society Bulletin 19: 137±142. RAINA, S. N. 1990. Genome organisation and evolution in the genus GARAY, L. A. 1979. The genus Phragmipedium. Orchid Digest July± Vicia. In S. Kawano [ed.], Biological approaches and evolutionary Aug: 133±148. trends in plants, 183±201. Academic Press, London. KARASAWA, K. 1979. Karyomorphological studies in Paphiopedilum, REICHENBACH, H. G. 1854. Xenia Orchidacea 1: 3. Orchidaceae. Bulletin of the Hiroshima Botanic Garden 2: 1±149. ROBERTSON, W. M. R. B. 1916. Chromosome studies. I. Taxonomic . 1980. Karyomorphological studies in Phragmipedium, Orchi- relationships shown in the chromosomes of Tettegidae and Acri- daceae. Bulletin of the Hiroshima Botanic Garden 3: 1±49. didiae: v-shaped chromosomes and their signi®cance in Acrididiae, . 1986. Karyomorphological studies on nine taxa of Paphiope- Locustidae, and Grillidae: chromosomes and variations. Journal of dilum. Bulletin of the Hiroshima Botanic Garden 8: 23±42. Morphology 27: 179±331. , AND M. AOYAMA. 1986. Karyomorphological studies on Cyp- ROLFE, R. A. 1896. The Cypripedium group. Orchid Review 4: 327± ripedium in Japan and Formosa. Bulletin of the Hiroshima Botanic 334, 363±367. Garden 31: 1±22. ROSSO, S. W. 1966. The vegetative anatomy of the Cypripedioideae , AND . 1988. Karyomorphological studies on two species (Orchidaceae). Botanical Journal of the Linnean Society 59: 309± of Paphiopedilum. Bulletin of the Hiroshima Botanic Garden 10: 341. 1±6. SESSIONS, S. K., AND A. LARSON. 1987. Developmental correlates of , AND K. SAITO. 1982. A revision of the genus Paphiopedilum genome size in plethodontid salamanders and their implications for (Orchidaceae). Bulletin of the Hiroshima Botanic Garden 5: 1±69. genome evolution. Evolution 41: 1239±1251. , AND R. TANAKA. 1980. C-banding study on centric ®ssion in SMYTH, D. R. 1991. Dispersed repeats in plant genomes. Chromosoma the chromosomes of Paphiopedilum. Cytologia 45: 97±102. 100: 355±359. , AND . 1981. A revision of chromosome numbers in STEBBINS, G. L. 1971. Chromosomal evolution in higher plants. Edward some hybrids of Paphiopedilum. Bulletin of the Hiroshima Botanic Arnold, London. Garden 4: 1±8. VOGT, S. 1990. Pollination in (Orchidaceae). Lin- KENTON, A., J. B. DICKIE,D.H.LANGTON, AND M. D. BENNETT. 1990. dleyana 5: 145±130. Nuclear DNA amount and karyotype symmetry in Cypella and WERNER, J. E., R. S. KOTA, AND B. S. GILL. 1992. Distribution of telo- Hesperoxiphion (Tigridieae; Iridaceae). Evolutionary Trends in meric repeats and their role in the healing of broken chromosome Plants 4: 59±69. ends in wheat. Genome 35: 844±848.