Repeated Evolution of Succulent

Home , Anacampseros, Didiereaceae, Echinocactus, Lepismium, Lewisiopsis, Portulacaria

Haseltonia 14: 26–36. 2008 26 HASELTONIA 14, 2008 27

Variations on a theme: repeated lineages of succulent life forms emerging from eage comprises about 2200 species, of which what was once recognized as the family Portu- about 90% are classified as succulents (Eggli evolution of succulent life forms in the lacaceae (Hershkovitz and Zimmer 1997; Ap- 2002; Anderson 2005; Hunt 2006). Relation- Portulacineae (Caryophyllales) plequist and Wallace 2001; Nyffeler 2007). In ships among the members of suborder Portu- addition to the cacti, there has been the evo- lacineae have recently been further clarified Reto Nyffeler lution of (1) perennial rosette-forming herbs (Nyffeler 2007). The family Portulaceae s.l. is with persistent succulent leaves (such as Lew- found to be paraphyletic; some of its species Institut für Systematische Botanik, Universität Zürich, isia in western North America; Fig 2), (2) pe- are more closely related to either Cactaceae or Zollikerstrasse 107, CH-8008 Zürich, Switzerland rennial herbs with fleshy roots and highly suc- Didiereaceae rather than to some other spe- culent or greatly reduced leaves (Anacampseros cies of the same family (Fig 13). Urs Eggli and Avonia; Figs 3, 4), (3) annual herbs with The genera Ceraria and Portulacaria are Sukkulenten-Sammlung Zürich, Mythenquai 88, CH-8002 Zürich, Switzerland prostrate, fleshy stems and succulent leaves most closely related to traditional Didierea- (including Portulaca oleracea, P. afra, and their ceae, while the genera Anacampseros, Avonia, Matt Ogburn & Erika Edwards relatives; Figs 5, 6), (4) small, pachycaulescent Grahamia s. l. (including Talinaria, Talinopsis, Department of Ecology and Evolutionary Biology, Brown shrubs or trees with succulent leaves (Ceraria and Xenia—in composite referred to as tribe University, 80 Waterman St, Providence, RI 02912, USA from southern Africa; Fig 7) and long shoot– Anacampseroteae; Rowley 1994, 1995), as short shoot differentiation (Didierea in Mad- well as Portulaca and Talinum (the latter in a agascar; Fig 8), (5) subwoody shrubs with tu- revised circumscription excluding the mem- berous roots and herbaceous leaves (Talinum bers of the former section Phemeranthus but paniculatum and relatives; Figs 9–11), and (6) including the Madagascan endemic genus Abstract: The succulent life form is a tried and true strategy for plants living in arid envi- slightly succulent scramblers or vines with a Talinella), form a well supported clade to- ronments. It has evolved in many distantly related lineages comprising 12,500 species from thickened stem base (such as Basella and re- gether with the family Cactaceae. This clade 70 flowering plant families and has spawned remarkable radiations. Three major groups are lated genera mainly from the tropics of the was referred to as the “ACPT clade” (that is, generally recognized: (1) stem succulents (that is, leafless cactus-like growth forms), (2) leaf New World; Fig 12). Anacampseroteae, Cactaceae, Portulaca, Tali- succulents, and (3) caudiciform and pachycaul succulents. All three lifeform groups are rep- This new phylogenetic picture allows us to num) by Nyffeler (2007). These findings re- resented in the relatively small suborder Portulacineae. Here we suggest that this diversity place the evolution of Cactaceae within a richer ceive good statistical support from molecular provides a unique opportunity to evaluate early cactus evolution within a richer contextual context and ask the following questions: (1) sequences of the chloroplast genome. Further- framework. We briefly review what we know about the phylogenetic relationships within the What do the cacti have in common with their more, morphological and anatomical charac- suborder Portulacineae (that is, Basellaceae, Cactaceae, Didiereaceae, and Portulacaceae) and other succulent relatives, and how do they dif- teristics provide additional evidence in favor the morphology and ecology of all major Portulacineae lineages. We then outline what we fer? (2) What did the ancestral Portulacineae of these inferred relationships (Ogburn 2007; believe to be key areas for future research on these understudied plants and discuss several look like, and where did it arise? (3) Can we Nyffeler and Eggli submitted). hypothetical “pre-adaptations” and conditions in ancestral Portulacineae that may have pro- infer the ecological conditions that may have Overall, we may recognize four major lin- moted the repeated evolution of unusual succulent life forms. triggered such dramatic morphological inno- eages in Portulacineae (Fig 13): (1) The resur- Key words: Caryophyllales, Portulacineae, Montiaceae, succulence, life form, functional vation in these lineages, and were they simi- rected family Montiaceae, here represented by trait, cactus, evolution lar in each case? (4) Can we infer anatomical the three genera Claytonia, Lewisia, and Phe- or functional pre-conditions that enable the meranthus, includes about 200 species mainly evolution of succulence? with herbaceous, rosettiform habits and clasp- Here we review what we currently know ing, non-constricted leaf bases. This family is Introduction Cactoideae and Opuntioideae is the result of about the evolutionary relationships, ecol- most prominent in the western parts of North Trained botanists and amateurs alike have re- the following modifications: (1) leaves highly ogy, and vegetative morphology of the Por- and South America. (2) The family Basella- garded the cacti with awe for centuries (Row- reduced, (2) short shoots (areoles) bear spines tulacineae and outline what we believe to be ceae forms a distinct and morphologically well ley 1997). The copious production of spines, derived from leaf primordia, (3) branching re- key areas for future research on these under- characterized clade of about 20 species that are lack of leaves, bizarre architecture due to the duced or absent, (4) long-lived stem epidermis studied plants. distributed in the tropics of the New World, formation of stem succulence, and the im- with delayed bark formation, and (5) cortex Africa, and Madagascar. (3) The family Didie- pressive ability to persist in warm arid des- and pith are expanded to form a water-stor- Phylogenetics of the reaceae is here used in the expanded circum- erts under some of the harshest environ- age tissue. Recent work on the phylogenetics suborder Portulacineae scription (Applequist and Wallace 2003) that mental conditions on Earth are all traits that of Pereskia (Edwards and others 2005; Butter- Today, the methods of molecular systematics includes Ceraria, Portulacaria, and possibly have entitled this lineage to be recognized as worth and Wallace 2005), as well as studies of allow us to reconstruct phylogenetic relation- also Calyptrotheca. These taxa generally form a prominent textbook example for adaptive Pereskia anatomy (Ogburn 2007), physiology, ships among groups of organisms in detail and large woody trees or shrubs, some with a dis- evolution in biology (for instance, Futuyma and ecology (Martin and Wallace 2000; Ed- with good measures of statistical support. Mo- tinct cactus-like habit, and occur in eastern 1997; Niklas 1997). Pereskia, a group of rel- wards 2006; Edwards and Donoghue 2006) lecular phylogenetic investigations of the past and southern Africa and Madagascar. (4) The atively non-succulent, leafy shrubs and small has revealed that this genus, indeed, has much decade (for instance, Hershkovitz and Zimmer ACPT clade consists of the family Cactaceae trees (Fig 1), has long been considered the to tell about early events in the evolution of 1997, 2000; Applequist and Wallace 2001; as well as three distinct subclades from the tra- “evolutionary link” between “ordinary” pe- the cactus life form (summarized in Edwards Cuénoud and others 2001) have clearly shown ditional family Portulacaceae. The genus Tali- rennial plants and leafless cacti (Rauh 1979; and Donoghue 2006). that the traditional families Basellaceae, Cac- num (including the genus Talinella, Nyffeler Gibson and Nobel 1986; Mauseth and Lan- Here we argue that there is even more to taceae, Didiereaceae, and Portulacaceae (that 2007) takes a cladistically basal position and drum 1997; but see Griffith 2004). The cac- gain by looking beyond Pereskia. It is now clear is, suborder Portulacineae) are closely related forms the sister group to a subclade consisting tus-form of representatives of the subfamilies that the cactus family is only one of several to each other. Overall, this evolutionary lin- of Anacampseroteae (Anacampseros, Avonia, and 28 Nyffeler AND OTHERS—evolution of succulent life froms HASELTONIA 14, 2008 29

Grahamia), Cactaceae, and Portulaca. This latter Distribution North, Central, and South Distribution North America (W half of the ground tuber, caudex or rhizome; leaves rosulate, group is well characterized by the presence of America, open areas, dry woodland to rocky continent only, SW Canada to NW Mexico), herbaceous, thin-textured to slightly succulent, axillary hairs or scales. The present hypothesis slopes. on rocks or gravel in usually open places. annually deciduous, flat to narrowly linear. (Fig 13; Nyffeler 2007) favors a sister-group re- Systematics and evolution Previously con- Systematics A genus of about 16 species Distribution North and Central Amer- lationship between Anacampseroteae and Cac- ceived as a subgenus of Talinum, but recent (Hershkovitz and Hogan 2002; Hershkovitz ica, northeast Asia (Mongolia, Siberia), usu- taceae, though this result only receives mod- molecular work (Hershkovitz and Zimmer and Hogan 2003). Lewisia tweedyi, formerly ally moist places (seepages, river banks, snow- erate statistical support. These findings have a 1997, 2000; Applequist and Wallace 2001; classified as Cistanthe, has now been segre- covered depressions) but some extending to profound bearing on the family classification Nyffeler 2007) has shown that this group of gated as monotypic genus Lewisiopsis. dry grassland and rocks. of the ACPT clade. If Cactaceae should fur- about 30 species is not at all related to Tali- Physiology Predominantly C3, facultative Systematics A genus of 26 species. Closely ther on be recognized in its familiar form, we num but belongs to the group of genera now CAM (inducible CAM-cycling) is reported by related to Montia (ca. 12 species, northern either need to propose additional families for recognized to belong to Montiaceae, where Guralnick and Jackson (2001). hemisphere circumboreal). the three subclades of former Portulacaceae, or it likely occupies a cladistically basal sister- Physiology Predominantly C3, faculta- we accept a recircumscribed paraphyletic Por- group position relative to the rest of the fam- Claytonia L. tive CAM (inducible CAM-cycling or mini- tulacaceae* that only includes Anacampsero- ily (Nyffeler 2007). The rather herbaceous, or Description Annual to perennial herbs usu- mal true CAM) is reported by Guralnick and teae, Portulaca, and Talinum. only slightly woody, rosulate habit is typical ally with fibrous roots; stem short, as under- Jackson (2001). for Montiaceae. Description of typical taxa of the Physiology Predominantly C3, facultative major Portulacineae lineages CAM (inducible CAM-cycling) is reported by Figure 5. Portulacaria afra. Figure 6. Portulacaria afra. Figure 7. Ceraria fruticulosa. Montiaceae Guralnick and Jackson (2001). Figure 8. Didierea madagascariensis. Figure 9. Talinum polygaloides. Phemeranthus Raf. Lewisia Pursh 5 6 Description Perennial; presumably long-lived Description Perennial long-lived herbs with herbs with taproots or underground tubers; fleshy taproot; stem none to short, contracted, stems usually none to short with contracted enlarged and forming a caudex; leaves rosu- internodes, herbaceous; leaves rosulate, her- late, herbaceous, slightly to strongly succu- baceous, terete, slightly succulent, annually lent, annually deciduous or long-lived, flat more or less deciduous (Fig 14). to terete (Fig 2).

Figure 1. Pereskia sacharosa (Bolivia). Figure 2. Lewisia rediviva (US). Figure 3. Anacampseros sp. Figure 4. Avonia papyracea. 1 2 7

3 4 9 30 Nyffeler AND OTHERS—evolution of succulent life froms HASELTONIA 14, 2008 31

Notes on other genera of Montiaceae flat to terete, succulent, annually deciduous, sally thickened and forming a caudex, some- and perhaps even characteristic. According to The three genera included in our survey do or short-lived evergreen (Fig 7). times upper parts deciduous; leaves alternate, Franz (1908), Portulaca is the only genus of not depict the whole range of life forms pres- Distribution South Africa (Namibia, slightly fleshy, deciduous to semi-deciduous Portulacineae showing Kranz anatomy, indi- ent in the family. Several taxa of Cistanthe RSA). (Figs 9, 10). cating C4 metabolism. show pronounced stem succulence, Hectorella Systematics and evolution A genus of seven Distribution Africa, Arabia, North and and Lyallia (previously recognized as a sep- species. Ceraria has traditionally been placed South America, some species pantropical but Grahamia Gill. s. str. (only G. bracte- arate family; Applequist and others 2006) in Portulacaceae and was only recently recog- perhaps introduced, seasonally mesic sites, semi- ata) are tight cushion plants occurring in alpine nized as a close relative of the formerly Mada- deciduous to deciduous forests and scrub. Description Perennial, presumably long- and sub-antarctic habitats (such as South gascar-endemic Didiereaceae (Applequist and Systematics and evolution A genus of 27 lived small shrubs with fleshy taproot; stems Island of New Zealand and Kerguelean Is- Wallace 2003). It forms a clade with Portu- (15 excluding Talinella) species. In traditional with long internodes, lignescent and not land), and Montiopsis species are small herbs lacaria (two species, leaves flat, Namibia and classifications Talinum also embraced Phe- succulent; leaves alternate, terete, succulent, to diminutive shrublets without any succu- South Africa), which is the sister-group to tra- meranthus, now removed to Montiaceae on annually ± deciduous or shortly persistent lence. The genera included in this resurrected ditional Didiereaceae. molecular grounds. On the other hand, the (Fig 16). family Montiaceae are representatives of the Physiology Leaves predominantly C3, Malagasy endemic, former genus Talinella Distribution South America (Argentina), so-called PAW clade (from Phemeranthus, CAM-cycling (Rundel and others 1999) or has been found to be nested within Talinum dry scrub (monte). Australian Calandrinia, western American inducible full-CAM (Guralnick and Jack- s. str. (Nyffeler 2007). Talinella species (ex- Systematics and distribution Recent mo- Portulacaceae; Hershkovitz 1993; Hershko- son 2001); stems with slightly higher CAM cluded from the above description) are small lecular systematic studies (Hershkovitz and vitz and Zimmer 2000). The systematics of activity (von Willert and others 1992; Gib- to large and sometimes lianoid shrubs. The Zimmer 1997; Nyffeler 2007) indicate that the groups dismantled from the former genus son 1996, Edwards and others in Winter and stems are hardly thickened, but at least one Grahamia s. lat. (Rowley 1994, 1995) does not Calandrinia (where Cistanthe and Montiopsis Smith 1996: 128). taxon produces a fleshy taproot and basal represent a monophyletic group but rather a have been traditionally placed) is still largely stem swelling, and the leaves of most taxa cladistically basal grade. The monotypic Gra- unresolved (Hershkovitz 2006). Didierea Baill. are slightly thickened. Furthermore, Tali- hamia s. str. and the morphologically similar Description Long-lived perennial shrubs or num s. str. and Talinella both often have monotypic genus Talinopsis from southern Basellaceae tree-like; stems woody, moderately succu- a dense cover of papillae on young stems United States (New Mexico, W Texas) and lent (medulla, cortex), eventually becoming (Figs 10, 11). neighboring Mexico form a basal grade within Basella L. woody, with contracted short-shoots in the Physiology Predominantly C3; inducible the tribe Anacampseroteae (Nyffeler 2007). Description Perennial vines with tuberous primary leaf axils; leaves linear, deciduous, CAM-cycling and weak full CAM (Guralnick Physiology No data available. underground roots and often basally swollen slightly succulent; short shoots first produc- and Jackson 2001). For Talinella we do not stems; stems slender, herbaceous, scrambling ing four spines, later producing a couple of have any data. Anacampseros Mill. s. lat. (incl. Avo- to twining; leaves alternate, flat, slightly fleshy, leaves each season (Fig 8). nia (Fenzl) G. D. Rowley) evergreen (Fig 12). Distribution South and southwest Mada- Portulaca L. Description Dwarf, perennial, presumably Distribution Madagascar, east Africa, one gascar, dry forest. Description Annual to perennial (but predom- long-lived herbs with tuberous roots or tap- species pantropical, usually in semi-deciduous Systematics A genus of two species. inantly short-lived) diminutive to conspicuous roots; stems usually contracted and short with to evergreen forests. Physiology No data available for Didierea, herbs or diminutive shrublets, with fibrous to short internodes, sometimes forming a swol- Systematics A genus of five species. but facultative CAM appears to be ubiquitous tuberous roots; stems herbaceous, slightly suc- len basal caudex with slender short-lived short Physiology C3 (Mabry 1977). in Didiereaceae s. str. (Kluge and Ting 1978; culent (rarely thin and lignescent, or strongly aerial shoots; leaves usually alternate and ar- Kluge and others 2001). succulent with flaking bark), internodes long ranged in rosettes, flat but much thickened Notes on other genera of Basellaceae to short; leaves alternate or opposite, flat to te- and strongly succulent, axils with hairs and/ This small family has only four genera and Notes on other genera of Didiereaceae rete, slightly to strongly succulent, not decid- or bristles or a rigid to papery scale. about 20 species. Almost all produce either In the traditional sense, the family embraced uous (but sometimes whole stems deciduous), Distribution South and east Africa, Austra- tuberous roots or thickened tuberous stem four genera with 11 species, all from Madagas- axils appearing naked but always with hardly lia; dry to very dry sandy to rocky places. bases and have weakly decumbent to twining, car, and with the exception of the monotypic noticeable to conspicuous tufts of hairs (rarely Systematics and evolution A genus of herbaceous stems and flat, slightly succulent Decarya (Fig 15) all show moderate stem suc- bristles or scales), rarely with a contracted short- about 25 species. Species with axillary scales leaves. The combination of spicate inflores- culence. Applequist and Wallace (2003) found shoot producing leaves (Figs 5, 6). are sometimes separated at generic level as cences and uniformly small and pale-colored in a molecular analysis that the South African Distribution Worldwide in tropical to tem- Avonia (Rowley 1994, 1995). Three species flowers also set the Basellaceae apart from genera Ceraria and Portulacaria (both weakly perate regions but with centers of diversity in of dwarf perennial herbs with rosulate suc- the rest of the Portulacineae. Some species of to pronouncedly stem-succulent shrubs) as South America, Africa, and Oceania, usually culent leaves from North and South America Anredera have functionally unisexual flowers well as the east African genus Calyptrotheca open arid to semi-arid sandy to rocky places, recognized either as monotypic genera (Tali- (Sperling and Bittrich 1993). (weakly leaf-succulent shrubs with or within scrub and deciduous forests, but also op- naria and Xenia; Gerbaulet 1992) or included out tuberous roots) belong here. portunistic and as weeds. in Grahamia s. lat. (Rowley 1994, 1995) are Didiereaceae Systematics A genus of 60–100 species of best included in an expanded genus Anacamp- ACPT Clade—Portulacaceae* unequivocal circumscription. seros s. lat. based on insights from molecular Ceraria Phearson & Stephens Physiology CAM appears to be widespread phylogenetic investigations and morphologi- Description Perennial woody shrublets to Talinum Adans. (Kluge and Ting 1978). Portulaca grandiflora cal reasoning (Nyffeler 2007). shrubs; stems thin and lignescent, or con- Description (Talinum s. str. excl. Talinella) shows inducible CAM in the stem and CAM- Physiology: C3 and inducible full CAM in tracted, much thickened and stubby, with Perennial herbs to small shrubs, usually with cycling in combination with C4 photosynthe- Anacamperos sect. Anacampseros (Guralnick and contracted invisible short-shoots in the pri- tuberous roots; stems elongate and lignescent sis in the leaves (Guralnick and others 2002). Jackson 2001); reports for Anacampseros sect. mary leaf axils; leaves alternate or opposite, to elongate and herbaceous, sometimes ba- Kranz anatomy is widespread (Nyananyo 1988) Avonia differ between exclusively C3 (Schütte 32 Nyffeler AND OTHERS—evolution of succulent life froms HASELTONIA 14, 2008 33

and others 1967) and inducible CAM (Run- Significant CAM-cycling has also been found OUTGROUPS del and others 1999). in the stems of some species (Martin and Wal- lace 2000). 0HEMERANTHUS ACPT Clade—Cactaceae Maihuenia Phil. ,EWISIA -ONTIACEAE Pereskia Mill. Description Perennial, long-lived, small cush- Description Perennial, long-lived shrubs to ion-forming shrublets with a fleshy taproot; #LAYTONIA small trees, roots fibrous or tuberous; stems stems short with short internodes, succulent with long internodes, slightly thickened and (medulla, cortex) becoming slightly woody "ASELLA "ASELLACEAE herbaceously succulent when young, soon with age, epidermis rapidly covered by a trans- becoming strongly woody and epidermis re- lucent periderm; leaves linear, terete, succu- placed by bark; leaves alternate, seasonally lent, seasonally mostly deciduous, axils with #ERARIA deciduous, flat, not or slightly fleshy; leaf an areole producing spines only. $IDIEREACEAE axils with a contracted short shoot forming Distribution South America (Argentina, $IDIEREA an indefinite number of spines and some- Chile); open gravelly to sandy places, cool times also seasonally deciduous leaves for climates. 4ALINUM many years. Systematics and evolution A genus of Distribution Southern Mexico to north- two species (Leuenberger 1997) that was pre- 0ORTULACA ern Argentina and Peru, Caribbean; decidu- viously included in subfamily Pereskioideae ous to semi-deciduous dry forest and open due to the lack of any obvious synapomor- 'RAHAMIA cactus scrub. phic characteristics. The succulent but hardly Systematics and evolution The genus (17 photosynthetic stems with stomata confined !NACAMPSEROS species; Leuenberger 1986) in its present cir- to the areolar pits are a remarkable develop- cumscription is paraphyletic and falls into ment in a relictual lineage from alpine habi- 0ERESKIA) !#04CLADE a northern clade (Pereskia I; Fig 13) and a tats (Mauseth 1999). southern clade (South America, Andean re- Physiology C3 with limited CAM-cycling 0ERESKIA)) gion) (Pereskia II) on molecular grounds (Ed- in both stems and leaves (Nobel and Hart- wards and others 2005; Butterworth and Ed- sock 1986; Mauseth 1999; Martin and Wal- -AIHUENIA wards, this volume, pp 46–53). The southern lace 2000). Pereskia clade is united by the presence of /PUNTIA stem stomata and a persistent green epider- Opuntia mis, which are traits they also share with other Description Perennial long-lived small shrubs %CHINOCACTUS cacti (the “core cacti,” consisting of Opunti- to large trees with fibrous or rarely fleshy roots; oideae, Cactoideae, and Maihuenia—though stems succulent, flattened, segmented, green Figure 13. Summary topology of 15 representatives of the suborder Portulacineae classified into three mono- it appears these traits were subsequently lost cladodes; leaves only as rapidly caducous, te- phyletic families (that is, Montiaceae, Basellaceae, Didiereaceae) and the ACPT clade. This latter subclade con- in Maihuenia) rete, green rudiments, axils with an areole pro- sists of a grade of three genera from traditional Portulacaceae and a monophyletic Cactaceae (based on Nyffeler Physiology CAM-cycling is present in ducing an indefinite number of spines and nu- 2007). Red bars mark representatives of traditional Portulacaceae (for instance, Carolin 1993). Dashed branches leaves of all studied species of Pereskia even merous glochids (minutely barbed, bristle-like indicate limited statistical support for the corresponding clade. under well-watered conditions. Drought-in- spines), often active for many years. duced full CAM has been recorded in leaves of Distribution Southwest Canada to cen- Mexico and the southern USA; usually open ered with bony hard aril, are diagnostic for Pereskia guamacho (Edwards and Diaz 2006). tral Argentina with a center of diversity in places or among deciduous to semi-decidu- the subfamily (except Pterocactus with flat- ous scrub or forest. tened winged seeds). The second diagnostic Systematics and evolution While some character for the subfamily is the presence Figure 10. Talinum triangulare. Figure 11. Talinella boiviniana. Figure 12. Basella excavata. authors treat the genus in a broad sense (that of glochids. 12 is, including the vast majority of the species Physiology Obligate full CAM appears to of subfamily Opuntioideae), Wallace and be ubiquitous in Opuntioideae (Gibson and Dickie (2002) have found that this results Nobel 1986; Nobel 1988; for lists of taxa see in a paraphyletic classification. In the nar- Gibson 1996: 118 and Sayed 2001). row sense, Opuntia (~180 species, still the largest genus of the family) is restricted to Echinocactus Link & Otto the “prickly pears,” with their typically flat- Description Perennial long-lived one- to sev- tened and jointed stem segments (aka clad- eral-headed stem-succulents; stem grossly en- odes or pads). In these “platyopuntias” only larged, barrel-shaped, with vertically oriented rudimentary and early caducous leaves are ribs; leaves as microscopic rudiments only; axils formed, but the genera Pereskiopsis and Quia- with a well-developed areole producing a small bentia of the same subfamily produce fleshy, but indefinite number of spines and then be- seasonally deciduous leaves as main photo- coming dormant; flowers solitary from the 10 11 12 synthetic organs. The seeds, which are cov- upper end of the areole, architecture of flower 34 Nyffeler AND OTHERS—evolution of succulent life froms HASELTONIA 14, 2008 35

and fruits as in Pereskia but ovary inferior; seeds 14 studies are needed), and they are also the only tion of succulent Anacampseros from the woody, brown to black, without strophiole. lineage to evolve obligate CAM photosynthe- thin-stemmed, mildly succulent Grahamia. Distribution North America (southern sis. They are also undeniably more “successful,” Is this a reversible trend? Currently, we don’t USA, Mexico), in open, semi-arid to highly whether this is measured by sheer ecological have much evidence of evolution occurring in arid rocky or gravelly places. abundance or species diversity. This may lead the opposite direction, though some species of Systematics A genus of six species, here one to infer that these traits confer some large Cactaceae (for instance, Disocactus, Epiphyllum, standing as representative of subfamily Cac- advantages over the other growth forms, but Lepismium, Rhipsalis) appear to have lost much toideae, which embraces about 80% of total there are many historical contingencies to con- of their succulence during evolution. We are cactus diversity. sider. Are the cacti much older than Didierea- only at the outset of investigating aspects of Physiology As for Opuntia. ceae or Montiaceae? Does biogeography and the adaptive strategy to evolve succulent shoots timing matter (for instance, were the cacti just under different environmental conditions (for The cacti: just another in the “right place at the right time”)? Dating instance, Mauseth 2000). weird Portulacineae the Portulacineae phylogeny, combined with Better phylogenetic resolution of some of Even without detailed analyses, several clear more rigorous analyses of historical biogeog- the less well known and more variable groups themes emerge when looking at the distribu- 15 raphy in this group, should provide us with (Talinum, Portulaca) will be critical for address- tion of traits across these lineages. (1) It ap- key information in this regard. ing this question (and critical also for more pears that virtually all Portulacineae exhibit accurate reconstructions of ancestral charac- some degree of succulence, whether in leaf, The road to succulence: ter states in Portulacineae). stem or root tissue. (2) Crassulacean acid me- a one-way street? In conclusion, we feel there is much to tabolism is prevalent as CAM-cycling, and it It is also tempting to view evolution in the Por- learn about early cactus evolution from gain- seems likely that all Portulacineae species have tulacineae as a repeated, inevitable transition ing a deeper understanding of the evolutionary some ability to at least recycle respired CO2 from mild succulence to extreme succulence. history of the cacti’s close relatives—a verita- during times of drought. (3) With the notable The evolution of the core cacti from woody, ble potpourri of enigmatic and woefully un- exception of some Montiaceae (that is, Clay- mildly succulent Pereskia-like ancestors, for derstudied succulents with different life form tonia, some Lewisia, Montia), most Portula- example, appears to be mirrored in the evolu- characteristics. cineae inhabit warm, drought-prone environ- 16 ments. Due to their ubiquitous distribution, it is highly probable that these traits are ple- Literature Cited siomorphic for the group and thus also char- Anderson EF. 2005. Das grosse Kakteenlexikon. Edwards EJ, Donoghue MJ. 2006. Pereskia and acterize ancestral Cactaceae (for similar ar- Ulmer Verlag. Stuttgart. (Augmented and the origin of the cactus life-form. American Nat- guments see Griffith 2004). This allows us to corrected German translation of the origi- uralist 167: 777–793. begin to develop a picture of the ecological nal English edition of 2001: The cactus fam- Edwards EJ, Nyffeler R, Donoghue MJ. 2005. and morphological conditions that preceded, ily. Timber Press. Portland.) Basal cactus phylogeny: implications of Pereskia and potentially promoted, the transition to Applequist WL, Wallace RS. 2001. Phylog- (Cactaceae) paraphyly for the transition to the the cactus life form in Portulacineae. eny of the portulacaceous cohort based on cactus life-form. American Journal of Botany Using anatomical, physiological, and envi- ndhF sequence data. Systematic Botany 26: 92: 1177–1188. 406–419. Eggli U, editor. 2002. Illustrated handbook of ronmental data from extant Pereskia popula- Applequist WL, Wallace RS. 2003. Expanded succulent plants: dicotyledons. Springer Verlag. tions, Edwards and Donoghue (2006) hypoth- Figure 14. Phemeranthus spinescens. Figure 15. De- circumscription of Didiereaceae and its divi- Berlin. esized that several key functional traits were carya madagascariensis. Figure 16. Grahamia bractea- sion into three subfamilies. Adansonia ser. 3 Franz E. 1908. Beiträge zur Kenntnis der Portu- present in cacti prior to the evolution of the 25: 13–16. lacaceen und Basellaceen. Botanisches Jahrbuch cactus life form, including (1) intolerance of seros lineage of the succulent Karoo and the Applequist WL, Wagner WL, Zimmer EA, für Systematik, Pflanzengeschiche und Pflanzen- large water deficits in tissues, (2) high photo- pachycaulescent Didiereaceae of Madagascar Nepokroeff M. 2006. Molecular evidence geographie 42 (Beiblatt 97): 1–46. synthetic water-use efficiency, (3) conservative and Africa evolved. This parallel evolution resolving the systematic position of Hec- Gerbaulet M. 1992. Die Gattung Anacampse- stomatal behavior with inducible CAM photo- of distinct succulent life-forms among three torella (Portulacaceae). Systematic Botany 31: ros L. (Portulacaceae). I. Untersuchungen zur synthesis, (4) distribution in warm, seasonally closely related lineages and from potentially 310–319. Systematik. Botanische Jahrbücher für System- Butterworth CA, Wallace RS. 2005. Molecular atik, Pflanzengeschiche und Pflanzengeographie dry habitats, (5) a shallow root system that can very similar ancestral conditions presents an phylogenetics of the leafy cactus genus Pereskia 113: 477–564. quickly absorb water from infrequent and un- unequaled opportunity for comparative stud- (Cactaceae). Systematic Botany 30: 800–808. Gibson AC. 1996. Structure-function relations of predictable rainfall events, and (6) some abil- ies of the different ways to “be” a succulent Cuénoud P, Savolainen V, Chatrou, LW, Pow- warm desert plants. Adaptations of desert organ- ity to buffer water deficits through the release plant. How do these groups compare in terms ell M, Grayer RJ, Chase MW. 2002. Mo- isms. Springer-Verlag. Berlin. of water stored in various tissues (that is, suc- of water storage capacity, drought tolerance, lecular phylogenetics of Caryophyllales based Gibson AC, Nobel PS. 1986. The cactus primer. culence). This model of the ancestral cactus is stomatal behavior, or growth rates? Do these on nuclear 18S rDNA and plastid rbcL, atpB, Harvard University Press. Cambridge, highly consistent with inferences (using the different succulent strategies occupy a simi- and matK DNA sequences. American Journal Massachusetts. above summaries of ecological and morpho- lar breadth of ecological space? At the outset, of Botany 89: 132–144. Griffith MP. 2004. What did the first cactus look Edwards EJ. 2006. Correlated evolution of stem like? An attempt to reconcile the morphological logical characteristics) of Portulacineae eco- the cacti stand apart from the others in sev- and leaf hydraulic traits in Pereskia (Cacta- and molecular evidence. Taxon 53: 493–499. logical function, suggesting that these traits eral important ways: they are the only Portu- ceae). New Phytologist 172: 479–489. Guralnick LC, Jackson MD. 2001. The occur- not only preceded the evolution of the cactus lacineae lineage to transfer all photosynthetic Edwards EJ, Diaz M. 2006. Ecological physiol- rence and phylogenetics of Crassulacean Acid lifeform, but may also have been the condition function to stem cortical tissue (with the po- ogy of Pereskia guamacho, a cactus with leaves. Metabolism in the Portulacaceae. International from which the highly specialized Anacamp- tential exception of Decarya, but functional Plant Cell and Environment 29: 247–256. Journal of Plant Sciences 162: 257–262. 36 Nyffeler AND OTHERS—evolution of succulent life froms

Guralnick LC, Edwards G, Ku MSB, Hock- etative anatomical characters in Cactaceae: the ema B, Franceschi VR. 2002. Photosynthetic genus Pereskia. Journal of Plant Research 110: and anatomical characteristics in the C4-crassu- 55–64. lacean acid metabolism-cycling plant, Portu- Nobel PS. 1988. Environmental biology of agaves laca grandiflora. Functional Plant Biology 29: and cacti. Cambridge University Press. Cam- 763–773. bridge, Massachusetts. Hershkovitz MA. 2006. Ribosomal chloroplast Nobel PS, Hartsock T. 1986. Leaf and stem CO2 DNA evidence for diversification of western uptake in the three subfamilies of the Cactaceae. American Portulacaceae in the Andean region. Plant Physiology 80: 913–917. Gayana Botanica 63: 13–74. Nyananyo BL. 1988. Leaf anatomical studies in Hershkovitz MA, Hogan SB. 2002. Lewisia. Il- the Portulacaceae (Centrospermae), with regards lustrated Handbook of succulent plants: dicotyle- to photosynthetic pathways. Folia Geobotanica dons. Eggli U, editor. Springer Verlag. Berlin. Phytotaxonomica 23: 99–101. pp 388–394. Nyffeler R. 2007. The closest relatives of cacti: in- Hershkovitz MA, Hogan SB. 2003. Lewisia. sights from phylogenetic analyses of chloroplast Flora of North America. FNA Editorial Com- and mitochondrial sequences with special empha- mittee. Oxford University Press. New York. 4: sis on relationships in the tribe Anacampseroteae. 476–485. American Journal of Botany 94: 89–101. Hershkovitz MA, Zimmer EA. 1997. On the Ogburn RM. 2007. Anatomical variation in Cac- evolutionary origins of the cacti. Taxon 46: taceae sensu lato. Master Thesis, University of 217–242. Missouri. St Louis. Hershkovitz MA, Zimmer EA. 2000. Ribo- Rauh W. 1979. Kakteen an ihren Standorten. Ver- somal DNA evidence and disjunctions of west- lag Paul Parey. Hamburg, Germany. ern American Portulacaceae. Molecular Phylo- Rowley GD. 1994. Anacampseros and allied genera. genetics and Evolution 15: 419–439. A reassessment. Bradleya 12: 105–112. Hunt DR, editor. 2006. The new cactus lexicon. Rowley GD. 1995. Anacampseros, Avonia, Graha- DH Books. Millborne Port. mia. A grower’ s handbook. British Cactus and Kluge M, Ting IP. 1978. Crassulacean Acid Me- Succulent Society. tabolism. Analysis of an Ecological Adaptation. Rowley GD. 1997. A History of Succulent Plants. In Billings WD, Golley F, Lage OL, Olson JS, Strawberry Press. Mill Valley, California. editors. Ecological Studies. Analysis and Synthe- Rundel PW, Esler KJ, Cowling RM. 1999. Eco- sis. Vol. 30. Springer-Verlag. Berlin. logical and phylogenetic patterns of carbon iso- Kluge M, Razanoelisoa B, Brulfert J. 2001. Im- tope discrimination in the winter-rainfall flora plications of genotypic diversity and phenotypic of the Richtersveld, South Africa. Plant Ecol- plasticity in the ecophysiological success of CAM ogy 142: 133–148. plants, examined by studies in the vegetation of Sayed OH. 2001. Crassulacen Acid Metabolism Madagascar. Plant Biology 3: 214–222. 1975–2000, a check list. Photosynthetica 39: Leuenberger BE. 1986. Pereskia (Cactaceae). 339–352. Memoirs of the New York Botanical Garden 41: Schütte KH, Steyn R, van der westhuizen 1–141. M. 1967. Crassulacean Acid Metabolism in South Leuenberger BE. 1997. Maihuenia—monograph African succulents: A preliminary investigation of a Patagonian genus of Cactaceae. Botanische into its occurrence in various families. Journal Jahrbücher für Systematik, Pflanzengeschiche und of South African Botany 33: 107–110. Pflanzengeographie 119: 1–92. Sperling CR, Bittrich V. 1993. Basellaceae. The Mabry TJ. 1977. The order Centrospermae.Annals families and genera of vascular plants. Kubitzki K, of the Missouri Botanical Garden 64: 210–220. editor. Springer Verlag. Berlin. 2: 143–146. Martin CE, Wallace RS. 2000. Photosynthetic von Willert DJ, Eller BM, Werger MJA, pathway variation in leafy members of two sub- Brinckmann E, Ihlenfeldt HD. 1992. Life families of the Cactaceae. International Journal strategies of succulents in desert with special refer- of Plant Sciences 161: 639–650. ence to the Namib desert. Cambridge University Mauseth JD. 1999. Anatomical adaptations to Press. Cambridge, Great Britain. xeric conditions in Maihuenia (Cactaceae), a Wallace RS, Dickie SL. 2002. Systematic impli- relictual, leaf-bearing cactus. Journal of Plant cations of chloroplast DNA sequence variation Research 112: 307–315. in the Opuntioideae. Studies in the Opuntioi- Mauseth JD. 2000. Theoretical aspects of surface- deae (Cactaceae). Hunt DR, Taylor NP, editors. to-volume ratios and water-storage capacities of David Hunt. Sherborne. 6: 9–24. succulent shoots. American Journal of Botany Winter K, Smith JAC, editors. 1996. Crassu- 87: 1107–1115. lacean Acid Metabolism. Biochemistry, ecophysiol- Mauseth JD, Landrum JV. 1997. Relictual veg- ogy and evolution. Springer Verlag. Berlin.