Amer. J. Bot. 76(5): 657-665. 1989.

CHROMOSOMENUMBERS IN '

GREGORY K. BROWN AND AMY JEAN GILMARTIN2,3 Departmentof Botany and Rocky Mountain Herbarium,University of Wyoming, Laramie,Wyoming 82071; and 2Departmentof Botany and Marion Ownbey Herbarium,Washington State University, Pullman, Washington99164

ABSTRACT Eighty-threechromosome counts are reportedfor 72 taxa of the Bromeliaceae.Fifty-eight of these counts are the firstknown chromosomenumber reports for their respectivetaxa. A model of chromosomalevolution in the Bromeliaceae(n = 25) is presented.The model is parsimonious and consistent with existing data on meiotic chromosome numbers within the family and in the closely related Velloziaceae (n = 9). Two hypothesizedpaleodiploids (n = 8 and n = 9) hybridizedto form a tetraploidthat in turn hybridizedwith the n = 8 lineage. The resultantn = 25 is the extant base number for the family. Two alternativehypotheses could explain the unique extant base number(n = 17) for Cryptanthus:1) Cryptanthusrepresents the paleotetra- ploid level, i.e., prior to the second round of hybridization,or 2) the lower numberrepresents the result of a more recent series of aneuploidreductions from n = 25. Given the existence of intergenerichybrids involving Cryptanthus,aneuploid reduction is the morelikely interpretation.

RECENT RESEARCH concerning Bromeliaceae ploidy, chromosome size bimodality, and the systematics and evolution (e.g., Brown and correlationof nonconcordancein meiotic and Gilmartin, 1984, 1986; Gilmartin and Brown, mitotic chromosome numbers with the epi- 1985, 1986b) has sparkedrenewed interest in phytic mode of growth. the study of Bromeliaceae chromosomes and The purpose of this paper is to describe re- chromosome evolution. Past chromosome sults of an ongoing meiotic chromosome num- number surveys in the family (i.e., Lindschau, ber survey within the Bromeliaceae, and es- 1933; Gauthe, 1965; Weiss, 1965; Sharmaand pecially subfamily . We also Ghosh, 1971; Till, 1984) have relied mostly presenta model for chromosome base number on mitotic material.The only major exception evolution for the family that is consistent and to this was Marchant (1967) who utilized parsimonious with existing data. meiotically active microsporocytes. There is great variability in reported mitotic chromo- METHODS AND MATERIALS- Floralbuds were some numbers (Brown and Gilmartin, 1986), collected in the field, or obtained from culti- and lack of concordance between mitotic and vated material at Marie Selby Botanical Gar- meiotic numbersfor some taxa within the fam- dens, Sarasota,Florida (SEL).Buds were fixed ily. This variability in mitotic number is re- in field-mixedFarmer's solution (100%EtOH: flected in the variable interpretationsof chro- glacial acetic acid; 3: l/v:v) to which a drop of mosome base numbers for the family. Brown saturatedaqueous ferric chloride (FeCl3-6H20) and Gilmartin (1986) summarized the pre- had been added. The latter enhances chro- vious controversy over base number deter- mosome stainability. After a minimum of 24 mination for Bromeliaceae,and discussed the hr, fixed buds are transferredto 70% EtOH. current level of knowledge concerning poly- See Gilmartin and Brown (1986a) for a com- plete description of the field collaboratornet- ' Received for publication 13 October 1987; revision work and its operation. accepted 28 October 1988. For chromosome squash preparations, in- We thank a dedicated group of field collaborators, with- dividual anthers were removed from the bud out whom this project would not have been possible: James Ackerman, Puerto Rico; Stephan Beck, ; Olga Be- in 70% EtOH and transferredto a pool of 1% navides, ; Elizabeth Bravo, ; David acetic carmine on a microscope slide. While in Brunner, ; I. Chacon, ; Hermes Cua- the stain, the anther is cut transverselyin half. dros, Colombia; Linda Escobar, Colombia; Gert Hatsch- Using ultrafine-tippedneedle and forceps, the bach, ; Stephen Koch, Mexico; Gustavo Martinelli, Brazil; Fernando Ortiz, Ecuador; Isidoro Sanchez Vega, sporogenous masses are squeezed from each ; and Rosa Subils, . Expert technical help microsporangiumthrough the median trans- was supplied by Carol Annable. We thank Ron Hartman, verse cut. The sporogenous masses are posi- Don Hauber, and two anonymous reviewers for their com- tioned toward the center of the stain pool and ments. This work was supported by collaborative research grants BSR-8607 187 (GKB) and BSR-8407573 (AJG) from a coverslip and gentle finger pressure are ap- the National Science Foundation. plied. The preparationis further flattened by 3 Deceased 10 February 1989. passingthe slide throughan alcohol flame sev-

657 658 AMERICAN JOURNAL OF BOTANY [Vol. 76 eral times. Heating the slide helps to rupture far homogeneous for the base number of x = the callose that encapsulates the microsporo- 25 (also see Brown and Gilmartin, 1986). All cyte. Squasheswere examined with phase con- repeatchromosome number reports made here trast microscopy, and documented using Ko- for Pitcairnioideae taxa corroborate one or dak Technical Pan 2415 and drawings. As more previous counts (i.e., Lindschau, 1933; standard practice, a minimum of five micro- Di Fulvio, 1967; Marchant, 1967; Brown et sporocytes with unambiguous meiotic figures al., 1984). With the addition of new genera (usually in diplotene, diakinesis, metaphase I, Brewcariaand Steyerbromeliapublished since or metaphase II) serve as the basis for chro- Smith and Downs (1974), Pitcairnioideaecon- mosome numberdetermination. In caseswhere tains 15 genera.Chromosome data are lacking a new chromosome number (e.g., for the followinggenera: Abromeitiella (2 spp.), leiboldiana) or abnormality (e.g., B-chromo- Ayensua (1 sp.), Brewcaria(1 sp.), Brocchinia somes or fragments)was encountered,as many (18 spp.), Connellia (4 spp.), Cottendorfia(24 as 18 unambiguous meiotic figureswere doc- spp.), Encholirium (12 spp.), Navia (74 spp.), umented. Voucher herbarium specimens are and Steyerbromelia(1 sp.). at WS unless otherwiseindicated (see Table 1). All graphicdocumentation of chromosomes is Tillandsioideae-With over 800 in at RM. The nomenclaturefollowed here is that six genera(Catopsis, 19 spp.; Glomeropitcairn- of Smith and Downs (1974, 1977, 1979). ia, 2 spp.; Guzmania, 126 spp.;Mezobromelia, 3 spp.; Tillandsia,410 spp.; ,250 spp.), RESULTS-Eighty-threechromosome counts this is the largestof the subfamilies. Published are reportedfor 72 taxa (Table 1). For the most chromosome number information now is part, reports are either the first known pub- available for all but Mezobromelia. The pri- lished chromosome number, or represent a mary focus of this research is Tillandsia, the previously unreported number for a taxon. largestgenus in the family. Representativesquash preparations are shown A comparison of published chromosome in Fig. 1-4. number data for Tillandsia (Brown and Gil- martin, 1986) has revealed a strikingdiscrep- -Approximately 15%of the ancy between mitotic (root tip) and meiotic ca. 570 specieswithin Bromelioideaeare known chromosome numbers. Prior attempts to dis- by at least one chromosome number report. cover trendsof chromosomalevolution within Unfortunately, these reports are not evenly Tillandsioideae(e.g., Lindschau,1933; Gauthe, spreadacross the 27 generarecognized by Smith 1965) had been hinderedby this variability in and Downs (1979). Most come from five gen- mitotic chromosome number reports (see era (, Billbergia, Cryptanthus,Neo- Brownand Gilmartin, 1986, for additionaldis- regelia, Nidularium), while 14 (51%) of the cussion). An explicit goal of our research has genera are unknown chromosomally (Andrea, been to determine the level of meiotic chro- 1 sp.; Androlepis, 1 sp.; Araeococcus, 5 spp.; mosome numbervariability within Tillandsia. Disteganthus, 2 spp.; Fascicularia, 5 spp.; The count of n = 22 for T. complanata(subg. Fernseea, 1 sp.; Greigia,26 spp.;Hohenbergia, Allardtia)differs from an earlierreport (Brown 40 spp.; Hohenbergiopsis,1 sp.; Neoglaziovia, et al., 1984; n = 20). Tillandsia complanata is 2 spp.; Ochagavia, 3 spp.; Orthophytum, 17 a wide ranging forest epiphyte (Greater An- spp.; Ronnbergia, 8 spp.; Wittrockia,7 spp.). tilles, CostaRica to Bolivia andN. Brazil;Smith Except for Cryptanthusand Aechmea til- and Downs, 1977) and may comprise cyto- landsioides (Martius ex Schultes f.) Baker, all geographicraces. These two counts correspond Bromelioideae genera thus far studied appear closely to the known northern and southern to have a meiotically established extant base limits for this species. number of x = 25 (see Brown and Gilmartin, The report for T. leiboldiana (Fig. 2) is the 1986). Cryptanthusis anomalous in having a first record of n = 19 within the family. This, base of x = 17 (Marchant, 1967) and the pos- and the T. complanata chromosome numbers sible significance of this is discussed later. presumably arose via dysploidy from an an- Aechmea tillandsioides (n = 21; Marchant, cestral n = 25. We view such chromosome 1967) would appear to be an aneuploid deriv- numbersas being derived. The two reportsfor ative. T. complanata and the remaining 13 reports for species of the subg.Allardtia presented here Pitcairnioideae-Chromosome counts are are the only known chromosome number re- available for six pitcairnioid genera (Deuter- ports for this subgenus. ocohnia, Dyckia, Fosterella, Hechtia, Pitcair- The chromosome number reports for T. lo- nia, Puya). Members of the subfamilyare thus rentziana and T. vernicosa (n = 25) are the May 1989] BROWN AND GILMARTIN-CHROMOSOME NUMBERS IN BROMELIACEAE 659

TABLE 1. Chromosomenumber reports: Bromeliaceae

Taxon n Voucher Bromelioideae aAechmeapurpureo-rosea (Hook.) 25 Brazil: Copuva Bahia Wawra Hatschbach48762 Ananas ananassoides(Baker) Lyman Paraguay: Dpto. Amambay B. Smith Brunner1488 var. ananassoides 25 Pitcaimioideae aDeuterocohniahaumanii Castell. 25 Argentina: Prov. Salta Subils 3556 D. longipetala(Baker) Mez 25 Argentina: Prov. Cordoba Di Fulvio & Subils 748 & 756 aDyckiamicrocalyx Baker Paraguay: Dpto. Paraguari var. microcalyx 25 Brunner1263 aD. tomentellaMez 25 Paraguay: Dpto. Presidente Hayes Brunner1346 aD. velascanaMez 25 Argentina: Prov. Salta Subils 3573 aFosterellarusbyi (Mez) 25 Bolivia: Dpto. La Paz Lyman B. Smith Beck 12076 aPitcairniabakeri (Andre) 25 Ecuador: Prov. Pichincha Andr6 ex Mez Bravo 843 aP. dendroideaAndre 25 Colombia: Dpto. Cauca Escobaret al. 4338 P. flammea Lindley Brazil: Edo. Rio de Janeiro var. roezlii (E. Morren) Lyman B. Smith 25 Martinelli11567 ap. poortmaniiAndr6 25 Ecuador: Prov. Zamora-Chinchipe Peters 200 ap. sceptrigeraMez 25 Ecuador: Prov. Pichincha Ortiz 837 aPuyaeryngioides Andr6 25 Ecuador: Prov. Loja Peters 002 ap. lilloi Castell. 25 Argentina: Prov. Salta Subils 3596 P. spathacea(Griseb.) Mez 25 Argentina: Prov. Cordoba Di Fulvio & Subils 751 Tillandsioideae aCatopsissessiliflora (Ruiz 25 Mexico: Edo. Puebla Lopez & Pav6n) Mez Koch & Wendt8513 aC. nitida (Hook.) Griseb. 25 Mexico: Edo. Oaxaca Koch 86124 aGuzmaniamitis Lyman B. Smith 25 Colombia: Dpto. Magdalena Cuadros3165 aG. multiflora(Andre) 25 Colombia: Dpto. Antioquia Andr6 ex Mez Escobaret al. 3944 Tillandsiasubg. Allardtia aT. adpressifloraMez 25 Ecuador: Prov. Zamora-Chinchipe Peters 1300 bT. complanataBenth. Costa Rica: Cartago Alto Coris subsp. complanata 22 Chacon 1754 aT. deppeanaSteudel 25 Mexico: Edo. Puebla Koch & Wendt859 ?T. dudleyiLyman B. Smith 25 Peru: Dpto. Cajamarca Sanchez Vega 4141 aT. fend/eri Griseb. 25 Peru: Dpto. Cajamarca Sanchez Vega3660 aT. incarnataKunth 25 Ecuador: Prov. Pichincha Bravo 817 T. latifolia Meyen avar. divaricata(Benth.) Mez 25 Peru: Dpto. Cajamarca Sanchez Vega 3815 var. latifolia 25 Peru: Dpto. La Libertad Sanchez Vega3863 aT. leibo/dianaSchldl. Mexico: Edo. Puebla var. leibo/diana 19 Koch & Wendt8514 660 AMERICAN JOURNAL OF BOTANY [Vol. 76

TABLE 1. Continued

Taxon n Voucher

?T. plumosa Baker 25 Mexico: Edo. Oaxaca Peterson& Annable4671 aT. pyramidataAndre 25 Ecuador: Prov. Pichincha Ortiz 834 aT. rubellaBaker 25 Colombia: Dpto. Magdalena Cuadros2020 aT. sigmoidea Lyman B. Smith 25 Colombia: Dpto. Magdalena Cuadros3160 aT. stenouraHarms Colombia: Dpto. Magdalena var. stenoura 25 Cuadros2016 aT. tovarensisMez 25 Peru: Dpto. Cajamarca Sanchez Vega 3679 25 Colombia: Dpto. Magdalena Cuadros3166 Subg. Anoplophytum aT. lorentzianaGriseb. 25 Argentina: Prov. Catamarca Subils 3661 aT. vernicosaBaker 25 Bolivia: Dpto. La Paz Beck 12072 Subg. Diaphoranthema aT. capillarisRuiz Lopez & Pav6n Argentina: Prov. Cordoba forma hieronymii (Mez) Lyman B. Smith 50 Di Fulvio & Subils 770 aT. recurvata(L.) L. 25 Ecuador: Prov. Pichincha Bravo816 aT. tricholepis Baker var. tricholepis 50 Argentina: Prov. Catamarca Subils 3682 50 Argentina: Prov. Salta Subils 3572 Subg. Phytarrhiza aT. dodsoniiLyman B. Smith 25 Colombia: Dpto. Nario Benavides4304 aT. duratii Vis. var. duratii 25 Bolivia: Dpto. La Paz Beck 11211 25 Argentina: Prov. Salta Palaci 547 aT. humilis C. Presl 25 Peru: Dpto. Cajamarca Sanchez Vega 3834 T. monadelpha(E. Morren) 25 Colombia: Mpio. Tumaco Baker Benavides6700 aT. purpureaRuiz Lopez & Pav6n 25 Peru: Dpto. Cajamarca Sanchez Vega 3666, 4140 aT. reichenbachiiBaker 25 Argentina: Prov. Salta Subils 3574 T. scaligeraMez & Sodiro 25 + 2-10 Ecuador: Rio Palenque fragments Dodson s.n. (SEL 81-822)

Subg. Pseudalcantarea T. viridiflora(Beer) Baker 25 Mexico: Edo. Oaxaca Koch 86123 Subg. Pseudo-Catopsis aT. diffusaLyman B. Smith 25 Peru: Dpto. Cajamarca Sanchez Vega3814 ?T. multifloraBenth. var. tomensis Lyman B. Smith 25 Peru: Dpto. Cajamarca Sanchez Vega 3864 aT. riocreuxiiAndre 25 Colombia: Mpio. Belco Escobar 7200 aT. ropalocarpaAndre 25 Peru: Dpto. Cajamarca Sanchez Vega3865 aT. subalata Andre 25 Ecuador: Prov. Pichincha Ortiz 8311 Subg. Tillandsia aT. baileyi Rose ex Small 25 U.S.A.: Texas Gardner1165 May 1989] BROWN AND GILMARTIN-CHROMOSOME NUMBERS IN BROMELIACEAE 661

TABLE 1. Continued

Taxon n Voucher aT. bourgaei Baker 25 Mexico: Edo. Sinaloa Peterson& Annable4156 aT. caput-medusaeE. Morren 25 El Salvador:Prov. Santa Ana Mazariego I T. fasciculata Sw. var. fasciculata 25 Costa Rica: San Jose Chacon 1751 25 Mexico: Edo. Chiapas Peterson& Annable4705 aT. filifolia Schldl. & Cham. 25 Mexico: Edo. Puebla Koch & Wendt854 T. imperialis E. Morren ex Mez 25 Mexico: Edo. Hidalgo Koch 842 T. juncea (Ruiz Lopez & Pav6n) 25 Mexico: Edo. Chiapas Poiret Martinez 17622 aT. kirchhoffianaWittm. 25 Mexico: Edo. Mexico Koch & Fryxell 8331 aT. mexicana Lyman B. Smith 25 Mexico: Edo. Oaxaca Peterson & Annable 4672 aT.polystachia (L.) L. 25 Mexico: Edo. Mexico Koch & Fryxell 8330 25 + Colombia: Dpto. Antioquia B chro- Escobaret al. 3659 mosomes aT. utriculataL. 25 West Indies: Monserrat Gilmartin2990 25 Puerto Rico: Dorado Ackerman& Montalov2004 aT. violaceaBaker 25 Mexico: Edo. Hidalgo Koch 8314 25 Mexico: Edo. Oaxaca Koch 86107 25 Mexico: Edo. Michoacan Peterson& Annable 4628 Vriesea elata (Baker) 25 Colombia: Dpto. Antioquia Lyman B. Smith Escobar et al. 3661 25 Colombia: Dpto. Guatape Escobar 7226 25 Colombia: Dpto. Magdalena Cuadros3167 aV. haematina Lyman B. Smith 25 Brazil: Edo. Rio de Janeiro Martinelli10644 aV. harmsiana (Lyman B. Smith) 25 Peru: Dpto. Cajamarca Lyman B. Smith Sanchez Vega 3656 aV. neoglutinosaMez 25 Brazil: Edo. Rio de Janeiro Martinelli11683 a V. psittacina (Hook.) Lindley var. psittacina 25 Brazil: Edo. Rio de Janeiro Martinelli11573 aV werkleanaMez 25 Mexico: Edo. Chiapas Peterson& Annable4708 aV. zamorensis (Lyman B. Smith) 25 Ecuador: Prov. Zamora-Chinchipe Lyman B. Smith Peters202 a First chromosome count for taxon. b New number for taxon. third and fourth counts for subgenusAnoplo- illaris (subg. Diaphoranthema).Till (1984) re- phytum.The earlierreports are for T. aeranthos ported mitotic "tetraploidlevel" numbers (2n (2n = 64; Gauthe, 1965, reported as T. dian- = 84-96) for this same species. thoidea Rossi) and T. tenuifolia L. (n = 25 + The reports for T. recurvataand T. tricho- 1-2 fragments;Marchant, 1967). lepis are the first meiotic counts for these For the first time, tetraploidy in Tillandsia species. They corroboratethe ploidy levels de- is documentedfrom meiotic materialin T. cap- terminedby Till ( 1984). An apparentpolyploid 662 AMERICAN JOURNAL OF BOTANY [Vol. 76

X A~ ;~~~~~4 5, IL #t*> XV

I~~~~~~~~~~~~~~~~S t- rv $ ,0 t I &~~~~~~~~~~~~~I 40'

t>r*~~~~~~~ kV *t 4 0

bivalents.~2VouheraritdinTbe1 Fi.14 Rersnttv meoi cho osm sqas prprtin fro Brmlaee Ba scl eul 5 gm t1. bivalents. Vouchers ar lted inTabl e '$

series exists in T. tricholepis where Till (1984) The count of n = 25 + 2-10 fragments for reported 2n = 50 (diploid) for variety macro- T. scaligera is the second for this species. This phylla and 2n = 90 and 94 ("tetraploid-level") report notes the variable number of chromo- for variety tricholepis. some fragments present at metaphase I. Our May 1989] BROWN AND GILMARTIN-CHROMOSOME NUMBERS IN BROMELIACEAE 663 initial report for T. scaligera (Brown et al., bromeliadchromosome number evolution that 1984) was published as n = c. 25. That report attempted to account for all numbers then was submitted in manuscriptas "n = 25 + 6- available. The scheme, hypothetically ances- 11 fragments at metaphase I," but editorial tral at x = 8, is necessarily complicated by policy dictated omission of this information. attempts to account for reportedmitotic num- It should be noted that the two T. scaligera bers (i.e., 2n = 32, 48, 56, 64) which do not collections came from different sites in Ecua- agree with the growing data for meiotic num- dor, collected during differentyears, and both bers.In this regard,the reportby Billings(1904) are maintained in cultivation at SEL. Study of of n = 16 for Tillandsia usneoides, the first additional T. scaligerapopulations is necessary publishedchromosome numberfor the family before attempting to interpretthe significance and one which has not yet been corroborated, of these fragments. appears to have lent an unfortunate bias to Tillandsia subgenera Pseudalcantarea (4 some subsequent chromosomal interpreta- spp.) and Pseudo-Catopsis(48 spp.) are poorly tions. The Billings (1904) reportwould suggest understood. The chromosome counts pre- a base number of 8 for the lineage. Curiously, sentedhere are the firstfor these two subgenera. a seriesof latermitotic studies(e.g., Lindschau, Within subg. Tillandsia,the meiotic number 1933; Gauthe, 1965; Weiss, 1965) report an report for T. fasciculata (n = 25) differs from abundance of numbers at the hexaploid, sep- two earlierreports. Gauthe (1965) reported2n taploid and octaploid levels based on x = 8 = 64 for T. fasciculata without varietal des- (i.e., 2n = 48, 56, 64). We should note that ignation, and 2n = 56 for variety venosispica chromosomalraces may be presentwithin this (reported as T. compressa Bert. ex Schultes). morphologicallyvariable and most widespread Tillandsiafasciculatais a polymorphicspecies species in the family (Smith, 1934). Billings that includes at least ten varieties (Smith and studied material collected in Louisiana, while Downs, 1977). Its geographicalrange includes for the T. usneoidescount made by Till Mexico, CentralAmerica, Florida, Caribbean (2n = 50; 1984) came from Argentinaor Bo- Islands, and northern . All va- livia. rieties are epiphytic, with two (laxispica and For the interpretationof chromosome num- venosispica)that are sometimes saxicolous. ber evolution, the use of somatically deter- A similar lack of concordance between mined (holdfast-roottips) chromosome num- meiotic and mitotic chromosome numbers is ber reports (e.g., Lindschau, 1933; Gauthe, encountered in T. imperialis. Gauthe (1965) 1965; Weiss, 1965; Sharmaand Ghosh, 1971) reporteda mitotic 2n = 64, while we report a that do not agreewith meiotically determined gametic number of n = 25. Tillandsia imper- numberswould seem to be an unwise practice. ialis typically is epiphytic in forests and may Whether apparent meiotic and mitotic chro- be saxicolous (Smith and Downs, 1977). mosome numbernonconcordance in epiphytic The first meiotic and diploid level count for bromeliads is real (e.g., unstable B-chromo- T. juncea is reported here. Both Lindschau somes), an artifact, or is the resultantcombi- (1933) and Gauthe (1965) reported(as T. jun- nation of bias and/or inaccurate observation cifolia) 2n = 96 for this species. by some previous workers has not yet been The first two chromosome number reports clearlyestablished. Research is now in progress for T. polystachia merit mention because of to examine this problem. At present, we feel B-chromosomes found in the Colombian col- that the most prudentapproach is to base our lections (Escobaret al., 3659). The number of interpretations and hypotheses about chro- B-chromosomes varied between floral buds mosome evolution for the Bromeliaceae on from an observed low number of two, to as meiotic chromosomenumbers (e.g., Marchant, many as six per cell. There were also indica- 1967; Brown and Gilmartin, 1983; Brown et tions that the B-chromosome number varied al., 1984; Varadarajanand Brown, 1985; the between anthers in the same bud. The collec- reports here). tion from Mexico (Koch and Fryxell 8330) Figure5 outlines a hypothesisto account for showed 25 bivalents without B-chromosomes. the extantBromeliaceae chromosome number, x = 25. The model is dibasic and involves DISCUSSION-The chromosome numbers hybridization and polyploidy of paleodiploid presentedhere continue to supportMarchant's base numbers x = 8 and 9 to yield a paleotet- (1967) proposalthat x = 25 is the base number raploid (n = 17). This was followed by hy- for the family. We view this as the extant base bridizationbetween a paleodiploid (n = 8) and number, since, according to criteria put forth the paleotetraploidlineage (n = 17) with poly- by Grant (1963, 1981), x = 25 is considered ploid stabilization at the hexaploid level (n = to be polyploid in origin. 25). The most recent common ancestorfor the McWilliams (1974) presented a scheme for three subfamilies was hexaploid. Electropho- 664 AMERICAN JOURNAL OF BOTANY [Vol. 76

(1985). Goldblatt and Poston (1988) have pro- posed a hypothesis for chromosome number evolution in the Velloziaceae that recognizes x = 9 as the paleodiploid base. Aware of an increasinglyapparent close (possibly sister tax- on) relationship between Bromeliaceae and Velloziaceae (Dahlgrenet al., 1985; Gilmartin and Brown, 1987; Menezes, personal com- munication; Frolick and Barthlott, 1987; Rankeret al., 1988), we speculatethat the pa- leodiploid n = 9 genome is shared by these two families. This notion is supported by re- 25 strictionfragment analyses of chloroplastDNA (Rankeret al., 1988). Chromosome numbers at the hypothesized paleodiploid level (i.e., n = 8 or 9) have never been reported for Bromeliaceae. Cryptanthus (Bromelioideae)is chromosomallyanomalous within the family at n = 17, and could, theo- retically,correspond to the paleotetraploidlevel within the model (Fig. 5). If further research were to support Cryptanthusas a true tetra- 1 7 ploid (i.e., 8 + 9 to 17) this would suggest its removal from subfamily Bromelioideae and the establishment of a new subfamily, Cryp- tanthioideae. However, an alternative, more likely hypothesis must be considered, where Cryptanthusis the productof decreasinganeu- ploidy from the extant base of x = 25. Such presumedproducts of aneuploid reductionare encountered in the Tillandsia (x = in T. umbellata Andre (n = 18; Brown et al., 1984) and in T. complanata (n = 20,22) and T. leiboldiana(n = 19), discussed earlier.Thus the "tetraploid-level"(n = 17) in Cryptanthus is very nearly attained in Tillandsia. Addi- 8 9 tional indirectsupport for the aneuploid origin Fig. 5. Proposedmodel for chromosomebase number of Cryptanthusrests in the fact that the genus evolution in Bromeliales.The extant base numbern = 25 exhibits numerous features considered to be is synapomorphicfor Bromeliaceaeand hypotheticallyde- derived within the subfamily (e.g., leaves di- rived from hybridizationand polyploidy involving a pa- morphic, in part;flowers polygamous, in part; leodiploid(n = 8) and paleotetraploid(n = 17).The dibasic stamens adnate to petals; few ovules; pollen paleotetraploiddeveloped from hypothesized hybridiza- 1-lept). Furthermore, reduction tion and polyploidy involving paleodiploidsn = 8 and n aneuploid ap- = 9. P, Pitcairnioideae;T, Tillandsioideae;B, Bromelioi- pears to have occurred in the large Brome- deae. lioideae genus Aechmea (175 spp.), and is pre- sumably responsible for the n = 21 report in A. tillandsioides(Marchant, 1967). At present, retic studies of two Tillandsia species by Soltis we view Cryptanthusas having evolved via et al. (1987) have shown these to be isozy- aneuploidy. mically diploid except for an additional iso- Intensive study of the reportedCryptanthus- zyme for PGM in T. recurvata. This electro- Billbergiaintergeneric hybrids (C. beuckeri x phoretic data suggests that "diploidization" of Billbergianutans and C. bahianus x Billbergia the dibasic paleohexaploid has occurred. nutans;Smith and Read, 1975)is needed.These Other hypotheses are possible, but this one hybrids were generated under horticultural is consistent with the data and existing notions conditions and have received little scientific (e.g., Goldblatt and Poston, 1988) concerning study beyond the nomenclaturework of Smith base numbers (i.e., x = 8 and/or 9) and chro- and Read (1975). Marchant (1967) has re- mosomal evolution within the superorder Bro- ported n = c21 for x Cryptbergia meadii meliiflorae of Dahlgren, Clifford, and Yeo (=Cryptanthusbeuckeri E. Morren x Billber- May 1989] BROWN AND GILMARTIN-CHROMOSOME NUMBERS IN BROMELIACEAE 665 gia nutans H. Wendl. ex Regel). A complete GOLDBLATT,P., ANDM. E. POSTON. 1988. Observations chromosomal characterizationof the parental on the chromosome cytology of Velloziaceae. Ann. Missouri Bot. Gard. 75: 192-195. species and hybrid individuals may provide GRANT,V. 1963. The origin of adaptations. Columbia more definitive clues as to the true chromo- University Press, New York. somal nature of Cryptanthus(i.e., ancient tet- . 1981. speciation. Columbia University raploid or aneuploid derivative). Press, New York. LINDSCHAU, M. 1933. Beitrage zur Zytologie der Bro- LITERATURECITED meliaceae. Planta 20: 506-530. MARCHANT.C. J. 1967. Chromosome evolution in the BILLINGS,F. W. 1904. A study of Tillandsia usneoides. Bromeliaceae. Kew Bull. 21: 161-168. Bot. Gaz. (Crawfordsville) 38: 99-121. MCWILLIAMS,E. 1974. Chromosome number and evo- BROWN,G. K., AND A. J. GILMARTIN. 1983. Chromo- lution. In L. Smith and J. Downs [eds.], Flora neo- somes of bromeliads. J. Bromeliad Soc. 33: 171-172. tropica monogr. (Pitcaimioideae) 14: 33-39. , AND . 1984. Stigma structure and variation RANKER,T. A., D. E. SOLTIS,P. S. SOLTIS,AND A. J. in Bromeliaceae-neglected taxonomic characters. GILMARTIN.1988. Molecular systematics of familial Brittonia 36: 364-374. and subfamilial relationships of the Bromeliaceae. , AND . 1986. Chromosomes of the Bro- Amer. J. Bot. 75, part 2: 200. meliaceae. Selbyana 9: 88-93. SHARMA,A. K., AND I. GHOSH. 1971. Cytotaxonomy of , G. S. VARADARAJAN, A. J. GILMARTIN, AND H. the family Bromeliaceae. Cytologia 36: 237-247. LUTHER. 1984. Chromosome number reports SMITH,L. B. 1934. Geographic evidence on the lines of LXXXV. Bromeliaceae. Taxon 33: 758-759. evolution in the Bromeliaceae. Bot. Jahrb. Syst. 66: DAHLGREN, R., H. T. CLIFFORD, AND P. F. YEO. 1985. 446-468. The families of the : structure, evo- , AND R. J. DowNs. 1974. Bromeliaceae (Pitcaim- lution and . Springer-Verlag, Berlin. ioideae). Flora Neotropica Monogr. 14: 1-658. Di FULVIO, T. E. 1967. Recuentos chromosomicos en , AND . 1977. Bromeliaceae (Tillandsioi- angiospermas . Kurtziana 4: 87-90. deae). Flora Neotropica Monogr. 14: 663-1492. FROLICK, D., AND W. BARTHLOTT. 1987. Micromor- , AND . 1979. Bromeliaceae (Bromelioi- phology of epicuticular waxes of monocotyledons. deae). Flora Neotropica Monogr. 14: 1493-2142. Abstracts, XIV International Botanical Congress, Ber- , AND R. W. READ. 1975. Notes on Bromeliaceae lin. XXXVI. Phytologia 30: 289-303. GAUTHt, J. 1965. ContributionA l'etude caryologique SOLTIS, D. E., A. J. GILMARTIN, L. RIESEBERG, AND S. des Tillandsi6es. MWm.Mus. Natl. Hist. Nat., Ser. B, GARDNER. 1987. Genetic variation in the epiphytes Bot. 16: 39-59. Tillandsiaionantha and T. recurvata(Bromeliaceae). GILMARTIN,A. J., AND G. K. BROWN. 1985. Cleistogamy Amer. J. Bot. 74: 531-537. in Tillandsia capillaris (Bromeliaceae). Biotropica 17: TILL, W. 1984. Sippendifferenzierung Innerhalb Til- 256-259. landsiasubgenus Diaphoranthema in Suldamerika mit , AND . 1986a. Bromeliaceae: an interna- Besonderer Beriucksichtigungdes Andenostrandes und tional cooperative research project. Taxon 35: 107- der Angrenzenden Gebiete. Ph.D. dissertation, Uni- 109. versity of Vienna, Austria. , AND . 1986b. Cladistic tests of hypotheses VARADARAJAN, G. S., AND G. K. BROWN. 1985. Chro- concerning evolution of xerophytes and mesophytes mosome number reports LXXXIX. Bromeliaceae. within Tillandsia subgenus Phytarrhiza (Bromeli- Taxon 34: 729. aceae). Amer. J. Bot. 73: 387-397. WEISS, H. E. 1965. Etude caryologique et cyto-taxono- , AND . 1987. Bromeliales, related mono- mique de quelques Bromeliacees. MWm. Mus. Natl. cots, and resolution of relationships among Brome- Hist. Nat., Ser. B, Bot. 16: 9-38. liaceae subfamilies. Syst. Bot. 12: 493-500.