sign in sign up
<<
Home , Aztekium, Cacteae, Epithelantha, Ferocactus macrodiscus, Geohintonia, Mammillaria albilanata

This article was downloaded by: [UNAM Ciudad Universitaria] On: 04 April 2013, At: 17:45 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Systematics and Biodiversity Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tsab20 Molecular phylogeny, origin and taxonomic implications of the tribe (Cactaceae) Monserrat Vázquez-Sánchez a , Teresa Terrazas a , Salvador Arias a & Helga Ochoterena a a Instituto de Biología, Universidad Nacional Autónoma de México, Apartado Postal 70–233, México, 04510, DF, México Version of record first published: 03 Apr 2013.

To cite this article: Monserrat Vázquez-Sánchez , Teresa Terrazas , Salvador Arias & Helga Ochoterena (2013): Molecular phylogeny, origin and taxonomic implications of the tribe Cacteae (Cactaceae), Systematics and Biodiversity, 11:1, 103-116 To link to this article: http://dx.doi.org/10.1080/14772000.2013.775191

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Systematics and Biodiversity (2013), 11(1): 103–116

Research Article Molecular phylogeny, origin and taxonomic implications of the tribe Cacteae (Cactaceae)

MONSERRAT VAZQUEZ-S´ ANCHEZ,´ TERESA TERRAZAS, SALVADOR ARIAS & HELGA OCHOTERENA Instituto de Biolog´ıa, Universidad Nacional Autonoma´ de Mexico,´ Apartado Postal 70–233, Mexico´ 04510, DF, Mexico´ (Received 13 October 2012; revised 25 January 2013; accepted 4 February 2013)

This study aimed to test the phylogenetic relationships of the tribe Cacteae, the generic circumscription within the tribe, in particular, the monophyly of the , and to provide a biogeographical hypothesis about the origin of Cacteae. The analysis included 135 from all of the 27 accepted genera and four outgroup species. Five chloroplast regions were sequenced, aligned, and coded postulating gaps, simple sequence repeats (SSRs), and inversions as potential synapomorphies, and their contributions to phylogenetic reconstruction were evaluated. The phylogenetic analyses recovered 63% of the genera as monophyletic. The contribution of rpl16, trnL-F and psbA to the phylogenetic signal was higher than in the two more slowly evolving genes (rbcL, matK), but the gaps and SSRs supported some of the genera. This result differs from those of previous phylogenetic studies in which less than 35% of the genera were recovered as monophyletic. In this work, and were re-circumscribed with five and four species, respectively. was found to be polyphyletic; 11 species correspond to Turbinicarpus s.str., whereas a highly supported clade corresponded to Rapicactus, and three species need further study. Contrary to its current circumscription, Ferocactus was not supported as monophyletic because it is polyphyletic concerning Glandulicactus, , and . We recognize this group of genera as the Ferocactus clade in which the species share the presence of scales in the pericarpel and ribbed stems, whether tuberculated or not. The Cacteae seem to have originated in the Sierra Madre Oriental and then dispersed to the Mexican Plateau, where radiation and diversification occurred at the boundaries of the Miocene–Pliocene Epoch. The development of the Mexican Plateau and the Trans-Mexican Volcanic Belt may have favoured the isolation of the Cacteae. A taxonomic diagnosis is presented for the tribe Cacteae and 18 genera that we now recognize. Key words: alignment coding, Ferocactus, Rapicactus, rpl16, , Turbinicarpus

Introduction at the stem tip and on the pericarpel and floral tube. The The recognition of the tribe Cacteae as a monophyletic second, , represents one of the largest genera of North American cacti (c. 150–180 spp.); they are low group is currently indisputable (Barcenas´ et al., 2011; growing, usually globose, and have distinctly tuberculate Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013 Hernandez-Hern´ andez´ et al., 2011); however, the limits between genera, the number of species within each genus, stems. Both of these genera were artificially delimited in and their relationships to one another remain matters of the main classification systems of the 19th century. For ex- debate (Barthlott & Hunt, 1993; Butterworth et al., 2002; ample, Schumann (1898) recognized only five genera in the present tribe Cacteae. These five were split into smaller Barcenas´ et al., 2011). For example, during the past decade, 24 to 27 genera (including 314 to 384 spp.) have been recog- genera in subsequent years based on the combination of structural stem characters (e.g. simple vs. shrubby; prune- nized (Table 1; Anderson, 2001; Guzman´ et al., 2003; Hunt et al., 2006). The Cacteae was based on two major genera ribs vs. prune-tubercles), flowers (e.g. presence or absence according to the taxonomic revision by Butterworth et al. of areoles, scales and/or trichomes), fruits (e.g. fleshy vs. (2002). The first, Echinocactus, has ribbed stems and rep- dry; type of dehiscence), and seeds (e.g. shape, colour, size, resents the most typical cacti of the deserts of Mexico and microrelief), until molecular methods were integrated with the USA. It is characterized by dense masses of trichomes phylogenetic studies (Metzing & Kiesling, 2008). Butterworth et al. (2002) provided the first phylogenetic Correspondence to: Monserrat Vazquez-S´ anchez.´ E-mail: analysis to test the monophyly of the tribe, including 26 [email protected] genera and 62 species, of which 15 belong in the genus

ISSN 1477-2000 print / 1478-0933 online C 2013 The Natural History Museum http://dx.doi.org/10.1080/14772000.2013.775191 104 M. Vazquez-S´ anchez´ et al.

Table 1. The number of genera recognized within the tribe Cacteae in the most recent classifications and this paper.

Anderson, 2001 (26 genera) Guzman´ et al., 2003 (28) Hunt et al., 2006 (24) This paper (18)

Acharagma Acharagma Acharagma (2 spp.) Ariocarpus Ariocarpus Ariocarpus (6) Astrophytum Astrophytum Astrophytum Astrophytum (5) Aztekium Aztekium Aztekium (2) Cochemiea Cochemiea Cochemiea (3) Coryphantha Coryphantha Cumarinia Cumarinia Cumarinia (1) Digitostigma Echinocactus Echinocactus Echinocactus Echinocactus (4) Echinomastus Epithelantha Epithelantha Epithelantha (1) Escobaria Escobaria Ferocactus Ferocactus Ferocactus Geohintonia Geohintonia Geohintonia (1) Glandulicactus Leuchtenbergia Leuchtenbergia Leuchtenbergia Leuchtenbergia (1) Lophophora Lophophora Lophophora (2) Mammillaria Mammillaria Mammillaria Mammillaria Mammilloydia Neolloydia Neolloydia Obregonia Obregonia Obregonia (1) Ortegocactus Ortegocactus Ortegocactus (1) Pediocactus Pelecyphora Pelecyphora Pelecyphora (2) Sclerocactus Sclerocactus Rapicactus (5) Stenocactus Stenocactus Stenocactus Strombocactus Strombocactus Strombocactus (2) Thelocactus Thelocactus Thelocactus Turbinicarpus Turbinicarpus Turbinicarpus Turbinicarpus (2)

Mammillaria. They used only the intron rpl16 and coded 2002; Hernandez-Hern´ andez´ et al., 2011) or were based insertions and deletions that were considered phylogeneti- on only a single molecular marker (Barcenas´ et al., 2011). cally informative. Nevertheless, they obtained a highly un- Therefore, a comprehensive study that includes many taxa resolved topology, most likely because of low character and and more than one molecular marker that aims to test the taxon sampling. With the exception of Mammillaria, which generic circumscription within the Cacteae, with emphasis has been the subject of phylogenetic studies (Butterworth on Ferocactus, is still lacking. & Wallace, 2004; Harpke & Peterson, 2006), the remain- The aim of this study was to test the monophyly of the der of the Cacteae genera with greater taxonomic diversity, genera within the tribe Cacteae, particularly emphasizing Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013 such as Coryphantha (c. 43 spp.), Escobaria (c. 20 spp.), the study of Ferocactus. For gaining a higher phylogenetic Ferocactus (c. 28 spp.), Stenocactus (c. 18 spp.), Thelo- resolution, we included a greater representation of each (c. 12 spp.) and Turbinicarpus (c. 21 spp.), have not taxon from each genus within the Cacteae, and we used been studied well enough to establish phylogenetic relation- five cp-DNA molecular markers (matK, rbcL, psbA-trnH, ships. Phylogenetic analyses have only been conducted in rpl16 and trnL-F) coding gaps, and SSRs (simple sequence Coryphantha (Zimmerman, 1985) and Sclerocactus (Porter repeats). Based on our results, we provide a taxonomic et al., 2000) based on structural and molecular data, synopsis for monophyletic or monotypic genera, discuss respectively. possible directions of further research for the genera that Ferocactus has diversified in contrasting environments were not monophyletic, and provide a hypothesis about the from the warm and dry regions of Mexico and the south- origin of Cacteae. western USA. In a phylogenetic study conducted by Cota & Wallace (1997) using chloroplast restriction site data and 15 species of Ferocactus, the genus appeared to be paraphyletic Materials and methods regarding F. glaucescens and F. histrix, which was sister to . More recently, Ferocactus was Terminal sampling also found to be non-monophyletic. However, these stud- The monophyly of the tribe Cacteae is well-supported ies had relatively poor taxa sampling (Butterworth et al., by previous studies based on morphological (Terrazas Molecular phylogeny of the tribe Cactae 105

& Arias, 2003) and molecular characters from chloro- denaturing the DNA for 2 min at 94 ◦C, followed by 29 plast DNA (Butterworth et al., 2002; Nyffeler, 2002). cycles of 1 min at 94 ◦C, 1 min at 48 ◦C, and 2.5 min at Therefore, we included four outgroup taxa with the sole 72 ◦C, and a final extension of 7 min at 72 ◦C. The matK purpose of tree orientation. Within the Cacteae, species exon was amplified using the primers matK 1R kim and were selected to include a representation of generic di- matK 3F kim (CBOL Working Group, 2009) by ini- versity based on different growth forms and floral mor- tially denaturing the DNA for 2.5 min at 94 ◦C, followed phologies. In total, 139 taxa, 135 species of the tribe by 30 cycles of 30 s at 94 ◦C, 40 s at 48 ◦C, and 40 s at Cacteae and 4 outgroups of the subfamilies , 72 ◦C, and a final extension of 5 min at 72 ◦C. The rpl16 in- Opuntioideae and Pereskioideae, were included. The sam- tron was amplified using the primers rpl161F and rpl163R pling represented 27 of 28 genera (Appendix S1, see (Hernandez-Hern´ andez´ et al., 2011) by initially denaturing supplementary material, which is available on the Sup- theDNAfor5minat94◦C, followed by 30 cycles of 1 min plementary tab of the article’s Taylor & Francis On- at 94 ◦C, 50 s at 55 ◦C, and 2 min at 72 ◦C, and a final line page at http://dx.doi/10.1080/14772000.2013.775191) extension of 4 min at 72 ◦C. The trnL-trnF region, which and more than 50% of the species of the Cacteae. The includes the trnL intron and spacer, was amplified using genus Pediocactus was not represented because of a lack the primers c and f designed by Taberlet et al. (1991). The of fresh tissue. Generic and species recognition follows PCR program used was 2 min at 94 ◦C; 30 cycles of 1 min Guzman´ et al. (2003). Because Mammillaria was the sub- at 94 ◦C, 50 s at 50–55 ◦C, and 1.5 min at 72 ◦C; and a ject of previous studies conducted with a wide sampling final extension of 7 min at 72 ◦C. The psbA-trnH intergenic of taxa (Butterworth & Wallace, 2004; Harpke & Peter- spacer was amplified using the primers psbA (Sang et al., son, 2006), we did not consider it necessary to include 1997) and trnH (Tate & Simpson, 2003) by initially dena- more than 11 representative species. The largest genera turing the DNA for 2 min at 94 ◦C, followed by 29 cycles were represented by five or more species, including the of 30 s at 94 ◦C, 30 s at 52 ◦C, and 1 min at 72 ◦C, and a type species when possible (Appendix S1, see supple- final extension of 7 min at 72 ◦C. mentary material, which is available on the Supplemen- The PCR products were sequenced at the Genomic tary tab of the article’s Taylor & Francis Online page at Center, Washington University (http://genome.wustl.edu/), http://dx.doi/10.1080/14772000.2013.775191). To test the without additional purification. All of the regions were se- monophyly of Ferocactus, 22 of 28 species and three sub- quenced using the amplification primers; only trnL-trnF species, as reported by Guzman´ et al. (2003), were included was sequenced using internal primers d and e (Taberlet in the analyses. et al., 1991). The assembly and editing of the sequences Tissue samples were taken from the epidermis of fresh was performed using Sequencher 4.7 (GeneCode). stems in native populations, with the exception of Fero- cactus flavovirens, which was obtained from the living col- lection of the UNAM Botanical Garden. The samples were Phylogenetic analysis immediately dried with silica gel. Seven sequences were ob- The sequences for each marker were manually aligned us- tained from GenBank for matK, psbA-trnH, rbcL(Pereskia ing WinClada (Nixon, 2002) following the criteria set by lychnidiflora AY875358, AY851568, AY875238) and rpl16 Ochoterena (2009). Microsatellites and hotspots, such as a (Acharagma aguirreana AF267915, Lophophora williamsii 302-bp region in psbA-trnH, were impossible to align and AF267943, P. lychnidiflora AY851594 and Strombocactus were excluded from the matrices. Upon alignment (avail- disciformis AF267967). able on request from the corresponding author), the gaps Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013 were coded using Simmons & Ochoterena’s (2000) simple method. The SSRs and inversions that were categorized as DNA extraction present or absent were coded manually in WinClada (Nixon, Total DNA was extracted using a DNeasy Kit (QIAGEN, 2002) following Ochoterena (2009). CA, USA) according to the manufacturer’sprotocol; except The regions sequenced in this study were considered to for a change in the tissue incubation time from 10 min to be inherited as one unit because they are from chloroplasts 1 h. Total DNA was stored at −20 ◦C. Five chloroplast (Doyle, 1992; Moore, 1995). Based on this principle, a markers, two exons, matK and rbcL, one intron, rpl16, single analysis was conducted using all five markers simul- and two intergenic spacers, psbA-trnH and trnL-F,were taneously. In addition, previous papers on cacti have eval- sequenced. Regions were amplified in a total volume of uated whether chloroplast markers have an identical origin 25 µL using the standard polymerase chain reaction (PCR, via the ILD test (Farris et al., 1994), providing congruent QIAGEN, CA, USA) protocol and contained 2.5 µLof10× topologies (Nyffeler, 2002; Butterworth & Wallace, 2004; buffer, 0.5 µLdNTPs,1µLBSA,0.5µL of each primer Arias et al., 2005). Because trnL-F was impossible to am- (10 mM), 0.75 µLofMgCl2, 0.125 µL Taq, and 19.125 µL plify for eight of the taxa (Cochemiea halei, C. pondii, C. of water plus 0.5–1 µLofDNA. poselgeri, Mammillaria albilanata subsp. tegelbergiana, The rbcL exon was amplified with rbcLa-F and rbcLa-R M. columbiana, Turbinicarpus boleanus, T. subterraneus primers (CBOL Plant Working Group, 2009) by initially and T. zaragozae), we used a strict consensus tree to detect 106 M. Vazquez-S´ anchez´ et al.

ambiguity (Wilkinson, 1995) and to conclude that the miss- used for the BBM analysis. The maximum number of ar- ing data did not create wildcard taxa; hence, we combined eas was kept as 6. The ancestral reconstruction is shown in the matrices despite the missing sequences. Fig. 2. The parsimony analysis consisted of 30 heuristic searches with 1000 random-addition-sequences, each using the four newly implemented technologies in TNT (Goloboff et al., Results 2008). All of the trees found were saved in every search Sequence length variation and the number of potentially to further eliminate sub-optimal and identical trees. The informative coded events for different markers are shown consensus tree was generated using the most parsimonious in Table 2. The topologies of the trees obtained in the con- trees resulting from all of the searches using the Nelsen op- sensus tree by parsimony and Bayesian analyses are largely tion (collapse and consensus) in WinClada (Nixon, 2002). congruent. Because identical major clades in the ingroup Clade support was estimated using jackknife and bootstrap were detected in both the Bayesian and parsimony analy- values with 1000 replicates, using an identical search strat- ses, we will use the parsimony tree to explain the phyloge- egy followed to find the most parsimonious trees (Goloboff netic relationships. The analysis resulted in a total of 372 et al., 2008). The values were calculated with respect to equally parsimonious trees with a step-length (L) of 3990, a the consensus of the most parsimonious trees. An unequiv- CI (Consistency index) of 0.33 (excluding non-informative ocal optimization of 100 characters (97 gaps, 2 SSRs and characters), and a RI (Retention index) of 0.68. The strict one inversion) was performed on the consensus tree using consensus tree (L:4038, CI:0.33, RI:0.67) with jackknife WinClada (Nixon, 2002) to explore their contribution to (JK) and bootstrap (BS) support values is shown in Fig. 1. clade support. The tribe Cacteae was found to be monophyletic (100% Bayesian analysis was conducted in MrBayes v. 3.2.1 BS/100% JK; 1 PP), defined from the unique combination (Huelsenbeck & Ronquist, 2001). The nucleotide substitu- of 42 characters, of which two were gaps, one was synapo- tion model employed six substitution types (nst = 6), with morphic (in trnL-F), and the other homoplastic (in trnH- base frequencies estimated from the molecular data. The psbA). Sixteen genera accepted by Guzman´ et al. (2003), or coding of gaps, SSRs and the inversion (using the Markov 12 genera based on Hunt et al. (2006), are also recognized model implemented in MrBayes), plus the five markers, in this study. were individually partitioned. The rate variation across Aztekium Boed. was sister to Geohintonia Glass & W.A. sites was modelled using a gamma distribution (rates = Fitz Maurice (100% BS/100% JK support, 1 PP). The sister invgamma). Bayesian posterior probabilities (PP) were cal- relationship (Aztekium–Geohintonia) was supported by 15 culated using four chains in the Markov chain Monte Carlo characters, two of them were gaps and one was synapo- (MCMC) analysis. The Bayesian analysis was based on two morphic (in rpl16). This clade was sister to the rest of the independent runs, and each analysis was run for 10,000,000 tribe. The clade that included four species of Echinocactus generations. The Markov chains were sampled every 1000 Link & Otto was defined by two gap characters (in trnL-F), generations. The burn-in was estimated by visual exam- one of which was a synapomorphy. The Astrophytum Lem. ination to assess convergence of the MCMC chains; the and Digitostigma Velazco-Macias & M. Nevarez´ sister re- remaining trees were summarised to calculate a major- lationship (100% BS/100% JK support, 1 PP) was defined ity rule consensus tree and the PP. All of the Bayesian by a unique combination of 19 characters, four of which analyses were performed in the freely available Bioportal were gaps, and two synapomorphies, one in rpl16 and the (http://bioportal.uio.no). other in psbA-trnH. The Astrophytum–Digitostigma plus Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013 Echinocactus clade has high support (92% BS/94% JK support, 1 PP) and was sister to the rest of the Cacteae. Origin of Cacteae Sclerocactus scheeri was sister to Echinomastus (100% The distribution range of Cacteae was divided into 13 areas BS/100% JK support, 1 PP), a clade that in turn was based on Morrone (2006). These areas are as follows: A sister to the remaining Cacteae, forming two main clades (California), B (Baja California), C (Sonora), D (Mexican (hereinafter called clades A and B). In the poorly supported Plateau), E (Tamaulipas), F (Sierra Madre Occidental), G clade A (47% BS/54% JK support), Echinocactus grusonii, (Sierra Madre Oriental), H (Trans-Mexican Volcanic Belt), three Ferocactus Britton & Rose species (F.haematacantus, I (Balsas Basin), J (Sierra Madre del Sur), K (Mexican F.glaucescens and F.histrix) and Leuchtenbergia principis Pacific Coast), L (Maracaibo) and M (Cauca). We used diverged early from the rest of the clade. The rest of the RASP (Yu et al., 2011) to perform a BBM (Bayesian bi- species were grouped into two subclades (A1 and A2). nary MCMC) analysis to reconstruct the possible ancestral Subclade A1 included representatives of Ferocactus, Glan- distribution of the tribe Cacteae on the phylogenetic tree. dulicactus (Backeb.) Buxb., and Stenocactus K.Schum. The MCMC chains were run simultaneously for 5,000,000 Subclade A2 included representatives of Ferocactus and generations, with 10 chains sampled every 100 genera- Thelocactus (K. Schum) Britton & Rose. Most of the tions. A fixed JC+G (Jukes-Cantor+Gamma) model was species in clade B have tubercles, whereas the ribbed Molecular phylogeny of the tribe Cactae 107 Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013

Fig. 1A. The strict consensus tree (L:4038, CI:0.33, RI:0.67) of 372 equally parsimonious trees with bootstrap (BS) and jackknife (JK) support values above branches and the number of characters supporting clades below. (∗) Relationships evident after gaps, SSRs and an inversion coded. Colours different than black represent the accepted genera.

species dominate in clade A. In clade B, which was mod- Ariocarpus Scheidw. and the rest of the Turbinicarpus erately supported, was the sister (Backeb.) Backeb. & Buxb. species than it was to the clade species to two large subclades (B1 and B2). In subclade formed by the first three diverging Turbunicarpus species. B1, the Strombocactus clade was more closely related to Although this Ariocarpus–Turbinicarpus relationship 108 M. Vazquez-S´ anchez´ et al. Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013

Fig. 1B. (Continued)

lacked support and was defined by only four characters, Obregonia–Lophophora (100% BS/100% JK support, the genus Ariocarpus was relatively well supported (88% defined by 28 characters) and Acharagma–Turbinicarpus BS/91% JK support, defined by 16 characters). In subclade (93% BS/100% JK support, defined by 13 characters). This B2, two main groups were formed (B2a and B2b). The B2a clade was sister to clade B2b (69% BS/87% JK support), clade (86% BS/96% JK support) was supported by 10 char- defined by a combination of 9 characters, including a acters, one of which was a gap (in psbA-trnH) and included synapomorphy in psbA-trnH. Neolloydia conoidea was Molecular phylogeny of the tribe Cactae 109 Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013

Fig. 2. Majority rule consensus tree resulting from the Bayesian analysis with five markers plus potentially informative events. Numbers at nodes are posterior probabilities. 110 M. Vazquez-S´ anchez´ et al.

Table 2. Summary of information obtained per marker.

matK rbcL rpl16 psbA/trnH trnL-F

No. taxa in the analysis 139 139 139 139 131 Non-aligned length variation 780–795 537–542 926–1237 323–647 796–1150 Length of aligned matrix 795 542 1458 878 1266 No. of potentially informative nucleotide sites 106 46 358 126 239 Coded events 7 6 160 53 90 Potentially informative events 3 4 66 31 41 Total characters used 109 50 424 157 280

sister to the remaining species in clade B2b, with low roborated as monophyletic (Appendix S2 for taxonomic support (69% BS/87% JK support). Cumarinia Buxb. synopsis of the monophyletic and monotypic genera, see and Cochemiea (K. Brandegee) Walton were sister taxa supplementary material, which is available on the Supple- and were defined by 8 characters, one of which was a mentary tab of the article’s Taylor & Francis Online page gap in rpl16. This clade was sister to the remaining taxa. at http://dx.doi/10.1080/14772000.2013.775191). This re- Escobaria Britton & Rose was found to be polyphyletic sult was in contrast to 30% and 20% of genera recovered with two species (E. chihuahuensis and E. laredoi) sharing as monophyletic by Butterworth et al. (2002) and Barcenas´ a common ancestor with Pelecyphora C. Ehrenb. (96% et al. (2011), respectively. We attribute the results presented BS/98% JK support) supported by 25 characters, one of here mainly to character sampling, not only to the inclusion which was a trnL-F gap, whereas the remainder two species of five chloroplast markers, but also to the coding of spe- (E. dasyacantha and E. missouriensis) formed a clade with cific putative evolutionary events (Ochoterena, 2009). The Echinomastus intertextus and . effect of including markers with different evolutionary rates This clade is part of a polytomy that includes several on the resolution of the tree was evident when the rbcL and genera. Ortegocactus Alexander has 7 autapomorphic matK exons were not included in the analyses, resulting in characters at markers rpl16 and psbA-trnH. Coryphantha a lack of a main phylogenetic backbone (results not shown). macromeris grouped with Echinomastus intertextus and Although these markers provide few potentially informa- two species of Escobaria in a weakly supported clade tive characters (159 combined, or nearly 16% of the total (Fig. 1). Five species of Coryphantha and Neolloydia number of potentially informative characters; Table 2), the matehualensis formed a strongly supported clade defined relationships among the main clades would not have been by 12 characters, two of which were homoplastic gaps (in evident without them. Simultaneously, faster evolving in- trnL-F, psbA-trnH), and one gap was synapomorphic (in tron and spacer regions (rpl16 intron, trnL-trnF intron and rpl16). Mammilloydia (Scheidw.) Buxb. was deeply nested spacer, and psbA-trnH intergenic spacer) provided support within Mammillaria Haworth, making the latter genus for the relationships among species and closely related gen- paraphyletic. This clade had low support and was defined era. Some of these relationships, nevertheless, were only by 11 characters, two of them homoplastic gaps (in trnL-F, evident after the gaps and SSRs were coded and included rpl16), and one was a synapomorphic SSR (in psbA-trnH). in the analysis (results not shown). Coding potential events, mainly gaps, was important for Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013 Origin of Cacteae better resolution of the phylogenetic relationships within the tribe Cacteae. Although numerous phylogenetic studies The BBM analysis supported the origin of Cacteae in the that include gaps are presently available, the results and Sierra Madre Oriental (SMO), with subsequent dispersal discussions of the benefits are limited (e.g. Simmons et al., into the Mexican Plateau (MP,Fig. 3). In the core Cacteae, at 2001; Ingvarsson et al., 2003; Ogden & Rosenberg, 2007). least five dispersal events from the Mexican Plateau to Baja The Simmons & Ochoterena (2000) simple gap coding California, California, Sierra Madre del Sur, Balsas Basin method used here is conservative and relatively easy to im- and the Trans-Mexican Volcanic Belt provinces might have plement. This method has been used with favourable results occurred. for phylogenetic reconstruction (e.g. Freudenstein & Chase, 2001; Hennequin et al., 2003; Ogden & Rosenberg, 2007). Discussion In the study presented here, other potential events, such as SSRs that were incorporated as potential synapomorphy The contribution of markers and the hypotheses (Ochoterena, 2009), also supported clades. coding of potential synapomorphies Moreover, results from the matrix analysis showed that In the present study, more than 63% of the 27 genera rec- the potential events that were coded were approximately ognized in Mexico by Guzman´ et al. (2003) were cor- 14% of the total data and supported the monophyly of Molecular phylogeny of the tribe Cactae 111 Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013

Fig. 3. Graphical result of origin at each node in the phylogeny of tribe Cacteae obtained by BBM analysis in RASP. Distribution range of Cacteae was divided into 13 areas based on Morrone (2006) as follow: A (California), B (Baja California), C (Sonora), D (Mexican Plateau), E (Tamaulipas), F (Sierra Madre Occidental), G (Sierra Madre Oriental), H (Trans-Mexican Volcanic Belt), I (Balsas Basin), J (Sierra Madre del Sur), K (Mexican Pacific Coast), L (Maracaibo) and M (Cauca). ∗ represents other ancestral ranges.

Acharagma (N.P. Taylor) Glass, Ariocarpus, Astrophytum, non-recovered sister taxa relationships in Barcenas´ et al. Aztekium, Cochemiea, Echinocactus, Lophophora J.M. (2011). The marker supporting the majority of clades was Coult., Mammillaria and Strombocactus Britton & Rose. rpl16, also used by Butterworth et al. (2002); however, Additionally, the events permit an understanding of the these studies did not consider gaps or SSRs, which in 112 M. Vazquez-S´ anchez´ et al.

our analyses, corresponded to 18% of the potentially accepted species (one according to Guzman´ et al., 2003; informative characters provided by this marker (Table 2). two according to Hunt et al., 2006), has tiny, numerous For the psbA/trnH region, almost 25% of the potentially tubercles, no larger than 1 mm, and numerous spines that informative characters corresponded to the gaps and SSRs, are appressed to the stem. In clade B1, the genus Strombo- whereas the trnL-F region included approximately 17% cactus, including the recently described species, S. corregi- of its potentially informative characters coded as gaps or dorae (Arias & Sanchez,´ 2010), was supported as mono- SSRs. We agree with Korotkova et al. (2011) that rpl16 is phyletic. This genus is characterized by its rhomboidal one of the best performing regions within the Cactaceae. tubercles in cross section, flowers borne at the plant tip, with few and tiny scales in the pericarpel, and very small seeds (Bravo-Hollis & Sanchez-Mejorada,´ 1991; Arias & Phylogenetic relationships Sanchez,´ 2010). The monophyly of Ariocarpus, including The early divergence of Aztekium and Geohintonia was con- 6 species, was concordant with previous studies using nar- sistent with previous reports by Butterworth et al. (2002). rower sampling or only a single marker (Butterworth et al., Other authors (Wallace, 1995; Butterworth et al., 2002; 2002; Barcenas´ et al., 2011; Hernandez-Hern´ andez´ et al., Hernandez-Hern´ andez´ et al., 2011) had previously noted 2011). Although two subspecies of A. retusus were recov- the relationship between Astrophytum and Echinocactus (E. ered as sister taxa, our results support raising A. retusus grusonii and E. polycephalus excluded). In our study, this subsp. trigonus to species level. This proposal is based on sister relationship is highly supported. Morphologically, the number of molecular autapomorphies, distinguishing both genera share the presence of numerous pointed scales A. retusus subsp. trigonus considerably from the type and trichomes covering the pericarpel and floral tube. Our species autonymous subspecies. In addition, the following results showed that core Echinocactus, with four species, morphological characteristics support the recognition of was supported as monophyletic (excluding E. grusonii,see both of the taxa as species: A. retusus has wide tubercles at section ‘Ferocactus’). Considering both the structural dif- the base, a tapered apex with spiniferous areoles, and white, ferences and previous phylogenetic studies, we propose the or rarely red, flowers, whereas A. trigonus K. Schum. pos- acceptance of Echinocactus as it is circumscribed here. sesses tubercles that curve downwards, yellow flowers and As previously suggested by Hunt (2003), our analy- no spiniferous areoles (Bravo-Hollis & Sanchez-Mejorada,´ ses strongly supported the transfer of Digitostigma caput- 1991; Anderson, 2001). Following the recent taxonomic medusae to Astrophytum. Astrophytum caput-medusae monograph of cacti (Hunt et al., 2006), Turbinicarpus has shares several characteristics with other species of Astro- been classified as a heterogeneous and superfluous group. phytum, such as floccose stems, flowers with pointed scales Some Turbinicarpus species have been frequently included and yellow perianth, fruit with regular dehiscence, per- in Bravocactus, Gymnocactus, Kadenicarpus, Neolloydia, sistent scales and trichomes, and hat-shaped seeds with Normanbokea or Pelecyphora (Anderson, 1986; Hunt & expanded testa that are inrolled around the sunken hilum Taylor, 1990; Barthlott & Hunt, 1993). Our results showed (Barthlott & Hunt 2000; Velazco & Nevarez,´ 2002). The that Turbinicarpus was polyphyletic and that its species remarkable and uncommon diagnostic character of A. were placed in three clades. Two clades were nested in caput-medusae (elongated tubercles) shows an interesting ‘clade B1’ and the other in ‘clade B2a’ (discussed below). convergence with Leuchtenbergia Hook., which suggests A core clade with 11 species, including seven subspecies, the need for an ontogenetic study to understand the devel- can be recognized as the genus Turbinicarpus s. str. because opment of tubercles. the type species T. schmiedickeanus is grouped here. This Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013 Sclerocactus Britton & Rose is distributed in the south- genus is characterized by small, globular and tuberculate ern USA, and there is independent evidence corroborat- stems, partial coverage by non-pungent ascending spines, ing that this genus is monophyletic (Porter et al., 2000; and tuberculate seeds with conjunct micropyles. A small Butterworth et al., 2002; Crozier, 2005). The sister rela- lineage, including T. horripilus and T. pseudomacrochele, tionship between Sclerocactus and Echinomastus Britton was found to diverge early in clade B1. Whether this & Rose was supported by a unique synapomorphy in psbA- clade is related to Pediocactus simpsonii, as suggested by trnH. Three species of Echinomastus were defined by an Butterworth et al. (2002), needs further investigation. exclusive set of rbcL, rpl16 and trnL-F characters. Thus, Groups from clade B2a included Obregonia Fric,ˇ further analyses are required to understand the natural cir- Lophophora, Acharagma and five species of Turbinicar- cumscription of Echinomastus because it could be poly- pus. The sister relationship of Lophophora and Obregonia phyletic. Hunt (1999) proposed the inclusion of Echino- is consistent with findings by Butterworth et al. (2002) and mastus in Sclerocactus; nevertheless, none of the available Hernandez-Hern´ andez´ et al. (2011). The two Lophophora phylogenies have sufficient sampling from either genus. species were found to be sister taxa, in contrast to results Our results suggested that Epithelantha F.A.C. Weber obtained by Butterworth et al. (2002). Our results allow ex Britton & Rose was sister to the remaining species of us to recognize Lophophora and Obregonia as separate clade B (Fig. 1). This genus, regardless of the number of genera, consistent with the conclusions of Anderson Molecular phylogeny of the tribe Cactae 113

(1967). Moreover, Lophophora has a papillose epidermis conclusion only for part of the species as in our sampling and no spines, whereas Obregonia has a non-papillose only two species are closely related to Coryphantha and epidermis and 3–4 spines. The two species of Acharagma Echinomastus. Hunt and Taylor (1990) suggested transfer- were also recovered as monophyletic and were sister to ring the monotypic genus Ortegocactus into Mammillaria. the third Turbinicarpus clade, consisting of five species (T. Using a wider sampling of Mammillaria, Butterworth & beguinii, T. booleanus, T. mandragora, T. subterraneous Wallace (2004) concluded that the inclusion of Ortegocac- and T. zaragozae). These results are consistent with tus within Mammillaria was inappropriate because of the Barcenas´ et al. (2011), so we propose the resurrection of likely polyphyletic nature of Mammillaria. However, our the genus Rapicactus Buxb. & Oheme ex Buxb., proposed analyses, which are supported by more molecular evidence, by Buxbaum (1942) and discussed by Luthy¨ (2003), for demonstrate that Ortegocactus could not be part of Mam- the five aforementioned species of Turbinicarpus. Species millaria as previously proposed by Hernandez-Hern´ andez´ of Rapicactus share medium-sized stems with high, dense, et al. (2011). straight and pungent spines, seeds that are striate, and with The genus Coryphantha was not supported as mono- the hilum and micropyle disjunct. phyletic. A group of five species plus Neolloydia mate- Clade B2b included 10 genera. Neolloydia was not mono- hualensis were defined by the combination of three molecu- phyletic according to our results. Neolloydia conoidea was lar events, whereas C. macromeris was more closely related the earliest diverging species of the clade. Butterworth & to Echinomastus intertextus and two Escobaria species. Ac- Wallace (2004) suggested that N. conoidea has a closer re- cording to Barcenas´ et al. (2011), based on a sample of 33 lationship to the Mammillaria species placed in the series species but only one molecular marker, Coryphantha is Ancistracanthae by Hunt (1981) and in the subgenus Co- not monophyletic, but a core Coryphantha consisting of 25 chemiea by Luthy¨ (2001). The inclusion of N. conoidea in species was recovered. Mammillaria was not supported here. The second species, The genus Mammillaria, as currently circumscribed N. matehualensis, appeared to be related to Coryphantha. sensu Butterworth et al. (2002) and Butterworth & Wallace New research that includes a denser sampling of Coryphan- (2004), is likely polyphyletic. Nevertheless, all the analyses tha will provide a stronger test to find out if N. mate- point to the existence of a core lineage of Mammillaria, hualensis is closely related to Coryphantha. The monotypic including the type species. Although our sampling was not genus Cumarinia was sister to Cochemiea. Cumarinia was exhaustive, our results support the recognition of Mammil- first described in the genus Coryphantha (Boedeker, 1930) laria, but it is still necessary to conduct an analysis with and then transferred to Neolloydia by Backeberg (1937). a more extensive sample of the genus to precise its limits. Buxbaum (1951) raised Cumarinia to generic rank. The tu- The core Mammillaria lineage seems to be paraphyletic in- bercle shape, spines, fruit and seeds show that the characters cluding Mammilloydia. Mammilloydia was raised to genus of Cumarinia are not like those of Neolloydia (Anderson, rank by Buxbaum (1950) because of differences in the seed 1986). Barcenas´ et al. (2011) considered Cumarinia as part testa with respect to Mammillaria. This proposal has been of the informal ‘Mammilloid clade’. However, this asser- previously accepted by several authors (Anderson, 2001; tion was not supported in this study when more molecular Guzman´ et al., 2003). The seed testa was analysed by Riha evidence was included because Cumarinia is more closely & Riha (1975), who concluded that it was similar to that related to Cochemiea than to Mammillaria. Nevertheless, of other Mammillaria species. Our results clearly support Cochemiea was proposed as a subgenus of Mammillaria those of Butterworth & Wallace (2004), who also suggested despite several differences in morphology (Butterworth & that Mammilloydia candida is a member of Mammillaria Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013 Wallace, 2004; Hunt et al., 2006; Barcenas´ et al., 2011). and recognized a core Mammillaria. In this study, we propose that Cochemiea should be re- As mentioned previously, until more species of garded as a genus separate from Mammillaria, which is Coryphantha and Escobaria are included in a phylogenetic supported by previous taxonomic treatments (Bravo-Hollis analysis, it would be mistaken to propose a circumscrip- &Sanchez-Mejorada,´ 1991; Barthlott & Hunt, 1993). Pele- tion for them. On the other hand, Pelecyphora is likely to cyphora was recovered as monophyletic as in other recent represent a lineage that probably will need to include some analyses (Barcenas´ et al., 2011). By contrast, Barcenas´ et al. Escobaria species. The circumscription of Pelecyphora will (2011) found Escobaria to be paraphyletic to Pelecyphora, likely change in the future; it is advisable to recognize the not polyphyletic as in our analysis. The species sampled in lineage by provisionally including the two accepted species both cases were different, showing the possibility that some within the genus. The case of Mammillaria is similar in that species of Escobaria may share a common ancestor with the preliminary results show the existence of a core group, Pelecyphora, while others not. Escobaria was considered which precise circumscription still requires research. How- part of Coryphantha (Berger, 1929) or closely related to it ever, the type of Mammilaria, M. mammillaris, is deeply by morphological (Taylor, 1986) and molecular evidence nested within the Mammilaria clade. Therefore, recognition (Butterworth et al., 2002). However, we would follow this of the genus is suggested. 114 M. Vazquez-S´ anchez´ et al.

Ferocactus demicity of the Cacteae (Hernandez´ & Barcenas,´ 1995; Our results showed that Ferocactus, as currently circum- Santa Anna del Conde-Juarez´ et al., 2009). During the scribed, was polyphyletic, which is consistent with previ- Cenozoic, extensional deformation was the main tectonic ous molecular phylogenies that included more taxa or that event that configured the MP in which large fault systems used different markers (Butterworth et al., 2002; Barcenas´ had a fundamental role in shaping the geologic evolution et al., 2011; Hernandez-Hern´ andez´ et al., 2011). We pro- of the region (Nieto-Samaniego et al., 2007). The final pose that all of the genera that are grouped in clade A events causing the uplift of the Sierra Madre Occiden- (Ferocactus, Glandulicactus, Leuchtenbergia, Stenocactus tal, which caused higher aridity in the MP, promoted the and Thelocactus) are considered Ferocactus until new evi- expansion of most of the succulent species, mainly the dence is incorporated. A unique character shared among all Cacteae. Afterward, the development of the Trans-Mexican of the species in clade A is the occurrence of scales in the Volcanic Belt (TVMB) in the Middle Miocene isolated pericarpel, and this feature is hypothesized to be a synapo- the MP from southern Mexico, providing conditions for morphy. Furthermore, the clade consists mainly of species the development of temperate and tropical communities with ribs, whether tuberculated or not. These genera were (Ferrusqu´ıa-Villafranca & Gonzalez-Guzm´ an,´ 2005), con- once considered to be members of the genus Ferocactus sisting of species such as Ferocactus macrodiscus and F. (Taylor 1979, 1980), with the exception of Leuchtenbergia, flavovirens, which are presently distributed in a tropical which has no evident morphological likeness with Ferocac- desert (Tehuacan Valley). With the uplift of the TMVB, tus because of its long triangular tubercles. Glass (1966) the humidity entering the MP was reduced, promoting suggested a Ferocactus–Leuchtenbergia relationship based higher aridity and the radiation of diverse succulent on their ability to produce hybrids in culture. An identical (Arakaki et al., 2011). Based on these data, it appears that relationship was later suggested by Barthlott & Hunt (1993) there were two major events in the origin of the Cacteae because of the presence of scales covering the pericarpel, tribe. The first event occurred 16 Ma and was related to the even though these scales are present in all of the species of origin of this group of succulent plants, and the second event this clade, as mentioned above. Butterworth et al. (2002) was a higher diversification within members of the Cacteae, and Crozier (2005) also found results that were consis- including Mammillaria, Coryphatha, Cochemia, Turbini- tent with those presented here, regarding the Ferocactus carpus and most of the members of Ferocactus clade, when and Glandulicactus relationship. The two Glandulicactus the aridity increased at the Miocene–Pliocene boundary. species included here are recognized as Sclerocactus sub- Our data also supported the radiation events associated with species by Hunt et al. (2006). However, our results showed the development of the Sonoran Desert during the Pliocene that Glandulicactus was not related to Sclerocactus. Steno- in northwestern Mexico, with the detachment of Baja Cal- cactus was considered a subgenus of Ferocactus by Taylor ifornia and the development of the Gulf. This event further (1980) that can be differentiated from other genera of the hindered the biotic exchange between the peninsula and clade by two characters: very thin and narrow ribs and are- the mainland, favouring the isolation of several taxa (Fer- oles with long spines in the upper part and small spines in rusqu´ıa-Villafranca & Gonzalez-Guzm´ an,´ 2005). Species the lower part. Nevertheless, Taylor (1980) noted characters from the Ferocactus clade and the genus Cochemiea pro- in S. coptonogonus that are shared with Ferocactus, involv- vide an example of isolation on the Baja California Penin- ing structures such as flowers, fruits and seeds, as well as sula. Nine species of Ferocactus sampled in this study are the presence of glandular spines. endemic to the Baja California Peninsula (Fig. 3, purple), Several Thelocactus species have also been included and they have relatives on the mainland; however, we do not Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013 within Ferocactus (Taylor, 1979) because of the presence know whether these species originated after the separation of extrafloral nectaries (Anderson & Ralston, 1978). The of the peninsula from mainland Mexico or what the direc- close relationship between both of the genera was corrob- tion and flow of movement was like between the mainland orated by our molecular study and may be supported after and the peninsula. the inclusion of vegetative and reproductive characters. In conclusion, our results prove that including more molecular evidence, considering potential synapomorphies such as gaps and SSRs, and the addition of more rep- Origin of Cacteae resentative species per genus, all support a greater reso- Our BBM analysis suggested that the Cacteae originated in lution of the relationships of the Cacteae that is consis- the Sierra Madre Oriental (SMO). The genera Aztekium and tent with recognized genera (see synopsis in Appendix S2, Geohintonia are distributed in some enclaves of the SMO, supplementary material, which is available on the Supple- which may support this hypothesis. The Cacteae experi- mentary tab of the article’s Taylor & Francis Online page enced upward shifts in the speciation rate (roughly 16–15 at http://dx.doi/10.1080/14772000.2013.775191). The re- Ma) based on the time-calibrated phylogeny of Arakaki cent diversification of the Cacteae on the Mexican Plateau et al. (2011). The SMO province and the Mexican Plateau and its dispersal to western Mexico (the provinces of Baja (MP) are the most important centres of diversity and en- California, California and Sonora) during the Pliocene is Molecular phylogeny of the tribe Cactae 115

consistent with the expansion of succulents worldwide, BARTHLOTT,W.&HUNT, D.R. 2000. Seed-diversity in the Cac- postulated by Arakaki et al. (2011). We believe that the taceae subfam. Cactoideae. Research 5, aforementioned geological events contributed to the de- 1–173. BERGER, A. 1929. Kakteen. Eugen Ulmer, Stuttgart, Germany. velopment of geographical barriers that favoured the iso- BOEDEKER, F. 1930. Eine neue, eigenartige Coryphantha. In: lation of the members of the Cacteae, contributing to its Monatsschrift. Deutschen Kakteen-Gesellschaft 2, 168–170. monophyly. Additionally, this study also shows the need BRAVO-HOLLIS,H.&SANCHEZ´ -MEJORADA, H. 1991. Las for several taxonomic changes, mainly to recircumscribe Cactaceas´ de Mexico,´ ed. 2. Vol. 2. Universidad Nacional Astrophytum, Echinocactus and Turbinicarpus. The inclu- Autonoma´ de Mexico,´ Mexico,´ D.F. BUTTERWORTH,C.A.&WALLACE, R.S. 2004. Phylogenetics stud- sion of additional evidence, e.g. vegetative and reproductive ies of Mammillaria (Cactaceae) insights from chloroplast se- characteristics, is expected to support our proposal. quence variation and hypothesis testing using the parametric bootstrap. American Journal of Botany 91, 1086–1098. BUTTERWORTH, C.A., COTA-SANCHEZ,J.H.&WALLACE,R.S. Acknowledgements 2002. Molecular systematics of tribe Cacteae (Cactaceae: Cactoideae): a phylogeny based on rpl16 intron sequence vari- We would like to thank the UNAM Graduate Program ation. Systematic Botany 27, 257–270. in Biological Sciences (Posgrado en Ciencias Biologicas,´ BUXBAUM, F. 1942. Rapicactus Buxb. et Oehme, Gen. nov. – Cac- UNAM) and National Council of Science and Technology taceae. Jahrbucher¨ der Deutschen Kakteen-Gesellschaft 1, 1–24. (Consejo Nacional de Ciencia y Tecnolog´ıa CONACyT) for BUXBAUM, F. 1950. The Morphology of Cacti. Abbey Garden the Ph.D. scholarship for MVS (41991). We appreciate the Press, Pasadera, CA, USA. support of Dennis Stevenson, who allowed us to initiate the BUXBAUM, F. 1951. Die phylogenie der nordamerikanischen molecular work of this project at the Molecular Laboratory Echinocacteen. Trib. Euechinocactineae F. Buxb. Oesterre-¨ at the New York Botanical Garden. Tania Hernandez´ pro- ichische. Botanische Zeitschrift 98, 44–104. CBOL PLANT WORKING GROUP. 2009. A DNA barcode for vided DNA from B. liliputana. Funding was provided by the land plants. Proceedings of the National Academy of Sciences Programa de Apoyo a Proyectos de Investigacion´ e Inno- USA 106, 12794–12797. vacion´ Tecnologica´ (Support Program for Research Projects COTA,J.H.&WALLACE, R.S. 1997. Chloroplast DNA evidence and Technological Innovation), DGAPA,UNAM IN224307 for divergence in Ferocactus and its relationships to North for TT and by UNAM Institute of Biology (Instituto de Bi- American columnar cacti (Cactaceae, Cactoideae). Systematic Botany 22, 529–542. olog´ıa). The authors thank two anonymous reviewers for CROZIER, B.S. 2005. Systematics of Cactaceae Juss.: Phylogeny, their useful comments to improve the manuscript. cpDNA Evolution, and Classification, with Emphasis on the Genus Mammillaria Haw. Dissertation, University of Texas, Austin, TX, USA. References DOYLE, J.J. 1992. Genes trees and species trees: molecular sys- ANDERSON, E.F. 1967. A study of the proposed genus Obregonia tematics as one-character taxonomy. Systematic Botany 17, (Cactaceae). American Journal of Botany 54, 897–903. 144–163. ANDERSON, E.F. 1986. A revision of the genus Neolloydia B. & R. FARRIS, J.S., KALLERSJO,M.,KLUGE,M.A.&BULT, C. 1994. Test- (Cactaceae). Bradleya 4, 1–28. ing significance of incongruence. Cladistics 10, 315–319. ANDERSON, E.F. 2001. The Cactus Family. Timber Press, Portland, FERRUSQU´IA-VILLAFRANCA,I.&GONZALEZ´ -GUZMAN´ , L.I. 2005. OR, USA. Northern Mexico’s Landscape, Part II: The biotic setting ANDERSON,E.F.&RALSTON, M. E. 1978. A study of Thelocactus across time. In: CARTRON, J.E., CEBALLOS,G.&FELGER, R.S., (Cactaceae) I. The status of the proposed genus Gymnocactus. Eds., Biodiversity, Ecosystems, and Conservation in Northern Cactus and Succulent Journal (Los Angeles) 50, 216–224. Mexico. Oxford University Press, New York, pp. 39–51. ARAKAKI,M.,PASCAL-ANTONIE,C.,NYFFELER,R.,LENDEL,A., FREUDENSTEIN,J.V.&CHASE, M.W. 2001. Analysis of mitochon- Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013 EGGLI,U.,OGBURN, R.M., SPRIGGS,E.,MOORE,M.J.& drial nad 1 b-c intron sequences in Orchidaceae: utility and EDWARDS J.E. 2011. Contemporaneous and recent radiations coding of length-change characters. Systematic Botany 26, of the world’s major succulent plant lineages. Proceedings of 643–657. the National Academy of Sciences USA 108, 8379–8384. GLASS, C. 1966. An incredible new bigeneric cultivar. Cactus and ARIAS,S.&SANCHEZ´ , E. 2010. Una especie nueva de Strombo- Succulent Journal (Los Angeles) 38, 177–178. cactus (Cactaceae) del r´ıo Moctezuma, Queretaro,´ Mexico.´ GOLOBOFF, P.A., FARRIS,J.S.&NIXON, K.C. 2008. TNT, a free Revista Mexicana de Biodiversidad 81, 619–624. program for phylogenetic analysis. Cladistics 24, 774–786. ARIAS,S.,TERRAZAS,T.,ARREOLA-NAVA , H.J., VAZQUEZ´ - GUZMAN´ ,U.,ARIAS,S.&DAVILA´ , P.2003. Catalogo´ de Cactaceas´ SANCHEZ´ ,M.&CAMERON, K.M. 2005. Phylogenetic relation- Mexicanas. Universidad Nacional Autonoma´ de Mexico´ y ship in Peniocereus (Cactaceae) inferred from plastid DNA Comision´ Nacional para el Conocimiento y Uso de la Biodi- sequence data. Journal of Plant Research 118, 317–328. versidad, Mexico´ D.F. BACKEBERG, C. 1937. Neolloydia Br. & R. Blatter¨ fur¨ Kakteen- HARPKE,D.&PETERSON, A. 2006. Non-concerted ITS evolution forschung 3: sine pag. in Mammillaria (Cactaceae). Molecular Phylogenetics and BARCENAS´ , R.T., YESSON,C.&HAWKINS, J.A. 2011. Molecular Evolution 41, 579–593. systematics of the Cactaceae. Cladistics 27, 470–489. HENNEQUIN,S.,EBIHARA,A.,ITO,M.,IWATSUKI,K.& BARTHLOTT,W.&HUNT, D.R. 1993. Cactaceae. In: KUBITZKI,K., DUBUISSON, J. 2003. Molecular systematics of the fern genus ROHWER,J.&BITTRICH,V.,Eds.,The Families and Genera Hymenophyllum s.l. (Hymenophyllaceae) based in chloro- of Vascular Plants, Vol. 2. Springer-Verlag, Berlin, Germany, plastic coding and no coding regions. Molecular Phyloge- pp. 161–197. netics and Evolution 27, 283–301. 116 M. Vazquez-S´ anchez´ et al.

HERNANDEZ´ ,H.M.&BARCENAS´ , R.T. 1995. Endangered cacti in OGDEN,T.H.&ROSENBERG, M.S. 2007. How should gaps be the Chihuahuan desert: I. Distribution patterns. Conservation treated in parsimony? A comparison of approaches using sim- Biology 9, 1176–1188. ulation. Molecular Phylogenetics and Evolution 42, 817–826. HERNANDEZ´ -HERNANDEZ´ ,T.,HERNANDEZ´ , M.H., DE-NOVA, J.A., PORTER, J.M., KINNEY,M.&HEIL, K.D. 2000. Relationships PUENTE,R.,EGUIARTE,L.E.&MAGALLON´ , S. 2011. Phylo- between Sclerocactus and Toumeya (Cactaceae) based on genetic relationships and evolution of growth form in Cac- chloroplast trnL-trnF sequences. Haseltonia 7, 8–19. taceae (, Eudicotyledonae). American Journal RIHA,J.&RIHA, J.J. 1975. The genus Mammilloydia Buxbaum. of Botany 98, 44–61. Cactus and Succulent Journal (Los Angeles) 47, 195–197. HUELSENBECK,J.P.&RONQUIST, F. 2001. MrBayes: Bayesian in- SANG,T.,CRAWFORD,D.&STUESSY, T.F. 1997. Chloroplast ference of phylogeny. Bioinformatics 17, 754–755. DNA phylogeny, reticulate evolution, and biogeography of HUNT, D. 1981. Revised classified list of the genus Mammillaria. Paeonia (Paeoniaceae). American Journal of Botany 84, Cactus and Succulent Journal of Great Britain 43, 41–48. 1120–1136. HUNT, D.R. 1999. CITES Cactaceae Checklist, 2nd edition. Royal SANTA ANNA DEL CONDE-JUAREZ´ ,H.,CONTRERAS-MEDINA,R.& Botanical Gardens and International Organization for Succu- LUNA-VEGA, I. 2009. Biogeographic analysis of endemic cacti lent Plant Study, Milborne Port, UK. of the Sierra Madre Oriental, Mexico. Biological Journal of HUNT, D.R. 2003. Astrophytum. Cactaceae Systematics Initiatives the Linnean Society 97, 373–389. 15, 5–6. SCHUMANN, K. 1898. Gesamtbeschreibung der Kakteen HUNT,D.R.&TAYL OR , N.P. 1990. The genera of Cactaceae: (Monographia Cactacearum). J. Neumann, Neudamm. progress toward consensus. Bradleya 8, 85–107. SIMMONS,M.P.&OCHOTORENA, H. 2000. Gaps as characters in HUNT, D.R., TAYL OR ,N.P.&GLASS, H. 2006. The New Cactus sequence-based phylogenetic analysis. Systematic Biology 49, Lexicon. Remous Ltd., Milborne Port, UK. 369–381. INGVARSSON, P.K., RIBSTEIN,S.&TAYL OR , D.R. 2003. Molecu- SIMMONS, M.P., OCHOTORENA,H.&CARR, T.G. 2001. Relative lar evolution of insertions and deletions in the chloroplast quality of gap characters in sequence-based phylogenetic anal- genome of Silene. Molecular Biology and Evolution 20, yses. Systematic Biology 50, 454–462. 1737–1740. TABERLET,P.,GIELLY,L.,PAUTOU,G.&BOUVET, J. 1991. Uni- KOROTKOVA,N.,BORSCH,T.,QUANDT,D.,TAYL OR , N.P., MULLER¨ , versal primers for amplification of three non-coding re- K.F.&BARTHLOTT, W. 2011. What does it take to resolve gions of chloroplast DNA. Plant Molecular Biology 17, relationships and to identify species with molecular mark- 1105–1109. ers? An example from the epiphytic Rhipsalideae (Cactaceae). TATE,J.A.&SIMPSON, B.B. 2003. Paraphyly of Tarasa (Mal- American Journal of Botany 98, 1549–1572. vaceae) and diverse origins of the polyploidy species. Sys- LUTHY¨ , J.M. 2001. A revised classification of the ‘primitive’ mam- tematic Botany 28, 723–737. millarias. Journal of Mammillaria Society 41, 6–7. TAYL OR , N.P. 1979. Notes in Ferocactus B.&R.Cactus and Suc- LUTHY¨ , J.M. 2003. Revision of the genus Rapicactus Buxbaum & culent Journal of Great Britain 41, 88–94. Oehme. Turbi-Now 14, 3–13. TAYL OR , N.P. 1980. Ferocactus and Stenocactus united. Cactus METZING,D.&KIESLING, R. 2008. The study of cactus evolution: and Succulent Journal of Great Britain 42, 108. the pre-DNA era. Haseltonia 14, 6–25. TAYL OR , N.P. 1986. The identification of Escobarias (Cactaceae). MOORE, W.S.1995. Inferring phylogenies from mtDNA variation: Cactus and Succulent Journal of Great Britain 4, 36–44. mitochondrial-gene trees versus nuclear-gene trees. Evolution TERRAZAS,T.&ARIAS, S. 2003. Comparative stem anatomy in the 49, 718–726. subfamily Cactoideae. Botanical Review 68, 444–473. MORRONE, J.J. 2006. Biogeographic areas and transition zones VELAZCO, M.C.G. & NEVAREZ´ , R.M. 2002. Nuevo genero´ de la of Latin America and the Caribbean islands based on panbio- familia Cactaceae en el estado de Nuevo Leon,´ Mexico:´ Dig- geographic and cladistic analyses of the entomofauna. Annual itostigma caput-medusae Velazco et Nevarez´ sp. nov. Cac- Review of Entomology 51, 467–494. taceas y Suculentas Mexicanas 47, 76–86. NIETO-SAMANIEGO, A.F., ALANIZ-A´ LVA R E Z ,S.A.&CAMPRUB´I,A. WALLACE, R.S. 1995. Molecular systematic study of the Cac- 2007. Mesa Central of Mexico:´ stratigraphy, structure, and taceae: using chloroplast DNA variation to elucidate cactus Cenozoic tectonic evolution. Geological Society of America phylogeny. Bradleya 13, 1–12. Special Papers 422, 41–70. WILKINSON, M. 1995. Coping with abundant missing entries in Downloaded by [UNAM Ciudad Universitaria] at 17:45 04 April 2013 NIXON, K.C. 2002. WinClada, version 1.00.08. published by the phylogenetic inference using parsimony. Systematic Biology author, Ithaca, New York [http://www.cladistics.com/, ac- 44, 501–514. cessed 6 February 2013]. YU,Y.,HARRIS,A.J.&HE, X.J. 2011. RASP (Reconstruct NYFFELER, R. 2002. Phylogenetic relationship in the cactus Ancestral State in Phylogenies) 2.0 beta [http://mnh.scu. family (Cactaceae) based on evidence from trnK/matK edu.cn/soft/blog/RASP, accessed 5 February 2013]. and trnL-trnF sequences. American Journal of Botany 89, ZIMMERMAN, A.D. 1985. Systematics of the genus Coryphantha 312–326. (Cactaceae). Dissertation, University of Texas, Austin, TX, OCHOTERENA, H. 2009. Homology in coding and non-coding DNA USA. sequences: a parsimony perspective. Plant Systematics and Evolution 282, 151–168. Associate Editor: Andreas Hemp