Demographic vulnerability in cliff-dwelling sonchus species endemic to the western mediterranean

1 Demographic vulnerability in cliff-dwelling Sonchus 1 species endemic to the western Mediterranean 2 Jose L. Silvaa,*, José A. Mejíasa & María B. Garcíab 3 aDepartment of Plant Biology and Ecology, University of Seville, Seville, Spain 4 bPyrenean Institute of Ecology, CSIC, Zaragoza, Spain 5 6 7 8 9 10 11 12 *Corresponding autor at: Departamento de Biología Vegetal y Ecología, 13 Universidad de Sevilla, Avenida Reina Mercedes s/n, 41012 Sevilla, Spain; Tel.: 14 +34 954556203. 15 E-mail adress: [email protected] (J.L. Silva). 16 17 Running title: Demographic vulnerability in cliff-dwelling Sonchus sect. 18 Pustulati. 19 *Manuscript 2 Abstract 20 Species of Sonchus section Pustulati (Sonchus masguindalii, S. fragilis and S. 21 pustulatus) constitute a group of endemic cliff plants in the Mediterranean 22 region, restricted to narrow non-overlapping areas within the Baetic-Rifan 23 hotspot of plant species diversity. S. pustulatus occurs both in SE Spain and N 24 Africa, whereas the other related species are exclusive to N Africa. We 25 characterized all the extant populations of Sonchus species (section Pustulati) 26 in 2008 by recording population size, demographic structure and reproductive 27 success, and we estimated the population trends for the critically endangered 28 Spanish S. pustulatus by repeating censuses in 2013. We also calculated the 29 stochastic population growth rate (λS) and modelled future viability (PVA) of one 30 of the Spanish S. pustulatus populations by using matrix models derived from 31 detailed demographic monitoring over a six-year period (2003–2008). 32 Population sizes ranged between 100-22000 reproductive individuals. In spite of 33 differences in plant size, population protection and anthropogenic disturbance, 34 the life history stage structure of all populations was similar and characterized 35 by a low abundance of seedlings and juveniles, suggesting low recruitment. The 36 population growth rate of Spanish S. pustulatus populations ranged between 37 0.91 and 1.01, and the matrix model showed significant population decline (λS = 38 0.9042; 95% CI: 0.9041–0.9043). The PVA projected that this population would 39 shrink to a few individuals in approximately forty years under present conditions. 40 Since high temperatures and drought negatively affected the dynamics of this 41 population, ongoing climatic change will jeopardize its future persistence. 42 43 3 Zusammenfassung 44 Die Sonchus -Arten der Sektion Pustulati (Sonchus masguindalii, S. fragilis and 45 S. pustulatus) bilden eine Gruppe endemischer Felspflanzen in der 46 Mittelmeerregion, die auf kleine, nicht überlappende Gebiete im Betikum/Rif-47 Hotspot der Pflanzendiversität beschränkt sind. S. pustulatus kommt sowohl in 48 Südost-Spanien als auch Nordafrika vor, während die anderen Arten der 49 Gruppe nur in Nordafrika auftreten. Wir beschrieben alle bestehenden 50 Populationen der Sektion Pustulati, indem wir im Jahre 2008 Populationsgröße, 51 demographische Struktur und Reproduktionserfolg erfassten, und wir 52 bestimmten die Populationsentwicklung für die kritisch gefährdete spanische S. 53 pustulatus, indem wir die Erhebungen 2013 wiederholten. Wir berechneten 54 auch die stochastische Populationswachstumsrate (λS) und modellierten die 55 zukünftige Überlebenswahrscheinlichkeit (PVA) für eine der spanischen 56 Populationen, auf der Basis von Matrix-Modellen für detaillierte 57 demographische Erhebungen über einen sechsjährigen Zeitraum (2003-2008). 58 Die Populationsgrößen variierten zwischen 100 und 22000 reproduktiven 59 Individuen. Trotz Unterschieden in der Pflanzengröße, im Schutzstatus und bei 60 anthropogenen Störungen war die Altersstruktur in allen Populationen ähnlich 61 und charakterisiert durch geringe Abundanzen von Sämlingen und 62 Jungpflanzen, was auf geringe Rekrutierung hinweist. Die Wachstumsraten der 63 spanischen S. pustulatus-Populationen rangierten zwischen 0.91 und 1.01, und 64 das Matrix-Modell zeigte eine signifikante Populationsabnahme (λS = 0.9042; 65 95% CI: 0.9041-0.9043). Die PVA ergab, dass diese Population bei gleich 66 bleibenden Bedingungen in ungefähr 40 Jahren auf wenige Individuen 67 zusammenschrumpfen wird. Da hohe Temperaturen und Trockenheit die 68 4 Entwicklung dieser Population negativ beeinflussen, wird der gegenwärtige 69 Klimawandel ihr Fortbestehen gefährden. 70 71 Keywords: Area of occupancy; Climatic trends; Population structure; PVA; Rare 72 plants; Sonchus section Pustulati; Stochastic population growth rate. 73 5 Introduction 74 For the conservation of rare or endangered species, a demographic approach 75 to species management is of critical importance (Schemske et al. 1994). 76 Detailed demographic characterisation in terms of population number, size and 77 structure across the distribution range of a species offers an accurate picture of 78 its current conservation status. Population monitoring can reveal both temporal 79 trends and the most critical life stages for the population growth rate (Horvitz & 80 Schemske 1995), but this is more time and resource consuming, and is 81 therefore often restricted to the most critically endangered species (Heywood & 82 Iriondo 2003). The study of both the current status and the dynamics of 83 populations are necessary to assess actual and potential threats and devise 84 appropriate management strategies. 85 Long-term monitoring programmes are uncommon because of the need to 86 maintain monitoring structures and funding in the long run. This becomes even 87 more difficult in the case of cliff-dwelling species due to their difficult access. 88 Consequently, our knowledge about the conservation and biology of cliff plants 89 is greatly limited (Larsson, Matthes, & Kelly 2005), despite the fact that rocky 90 habitats harbour many rare and endangered plants with narrow distributions; 91 particularly in the Mediterranean region (Thompson 2005). Cliffs are highly 92 fragmented, and offer very particular ecological environments that make life 93 difficult, such as scarce microsite availability for plant establishment. However, 94 competition in these habitats is also lower, and adapted organisms have a lower 95 probability of suffering anthropogenic disturbances. 96 (Larsson 6 Although limited, studies on the populations and dynamics of cliff-dwelling 97 plants suggest that local persistence is high (Lavergne, Thompson, Garnier, & 98 Debussche 2004; Thompson 2005), they are extraordinarily long-lived (Larson 99 et al. 2000; García, Guzman, & Goñi 2002), and their population sizes are 100 unusually stable (Morris & Doak 1998; Picó & Riba 2002; García 2003). 101 In this study, we examine the demography of a small group of related plants 102 that constitute the subgenus Sonchus section Pustulati (Asteraceae): Sonchus 103 pustulatus Willk., S. fragilis Ball, and S. masguindalii Pau and Font Quer 104 (Boulos 1973; Kim, Lee, & Mejías 2007). These narrow endemics are restricted 105 to small areas of the western Mediterranean Basin (Boulos 1973) on both sides 106 of the Alboran Sea (Fig. 1) and tend to be located on rocky cliffs at low altitudes. 107 All three species are found in North Africa, where they are considered to be 108 very rare (Fennane & Ibn Tatoo 1998). S. pustulatus also occurs in the SE 109 Iberian Peninsula, where it is categorized as “critically endangered” (Cueto et al. 110 2003). Phylogenetic and phylogeographic data suggest that these taxa are 111 relicts of the late Tertiary period (Silva 2014). Their restricted and disjunct 112 distribution seems to be related to old geological events of large biogeographic 113 impact: the Messinian Salinity Crisis (during which North Africa and the Iberian 114 Peninsula were connected by land), and the subsequent Zanclean reflooding of 115 the Mediterranean basin (by which Africa and the Iberian Peninsula were 116 separated). 117 The principal goal of this study was to assess the current conservation status 118 and future risk of extinction for these cliff-dwelling species and their vulnerability 119 to ongoing climate change. For that purpose, we undertook extensive fieldwork 120 7 to analyse the overall demographic situation of the clade, and modelled the 121 dynamics of one of the Iberian populations in detail. We focused on the 122 following specific objectives: (1) to accurately define the boundaries of the 123 distribution range of the species after visiting all known populations and 124 potentially suitable habitats; (2) to estimate extent of occurrence of each taxon, 125 as well as area of occupancy, size, structure, and reproductive success of all 126 extant populations; (3) to analyze temporal trends in the three Spanish 127 populations of S. pustulatus, and to model the viability (PVA) of the a priori most 128 endangered population by matrix models. Our results will help in assessing the 129 conservation status and vulnerability of a group of cliff-dwelling species of high 130 biogeographic and conservation value. 131 132 Material and methods 133 Plant species 134 The subgenus Sonchus section Pustulati includes three species: Sonchus 135 pustulatus, S. fragilis and S. masguindalii. They are restricted to four non-136 overlapping small areas within the Baetic-Rifan complex biodiversity hotspot 137 (western Mediterranean; Fig. 1, A-D), always occurring on low cliffs located on 138 the coast or less than 20 km from the sea. Some herbarium specimens of S. 139 pustulatus collected in the 19th century (e.g., herbarium COI) indicate that it was 140 also present in rocky places on the oceanfront near Ghazaouet (Fig. 1, E, 141 northwestern Algeria), but we were unable to visit this area and their presence 142 at this site could therefore not be confirmed. 143 8 The Sonchus section Pustulati includes suffrutescent chamaephyte plants, with 144 flower heads comprising approximately 30–120 florets in the case of S. 145 pustulatus and S. fragilis, and 60–250 florets in S. masguindalii. They bloom in 146 the spring and are visited by many species of Diptera and Coleoptera. The 147 incidence of self-incompatibility that enforces outcrossing is high in S. 148 pustulatus and S. masguindalii, and very low in S. fragilis (Silva 2014). Fruits 149 are achenes with a short-lasting pappus, released during late spring and early 150 summer. 151 152 Distribution area, extent of occurrence and area of occupancy 153 We conducted an intensive field work campaign from spring to early summer in 154 2007 and 2008. We visited all the locations for each of the three species that we 155 found recorded in different herbariums (BC, BM, COI, G, HUAL, MA, MPU, 156 RNG and SEV) and bibliographic references (Boulos L. 1973; Deil & Galán de 157 Mera 1996; Deil & Hammoumi 1997; Cueto et al. 2003; Mota et al. 2005), plus 158 potentially suitable habitats nearby. Following the IUCN criteria (IUCN 2001), 159 the extent of occurrence for each taxon was estimated as the area contained 160 within the shortest continuous imaginary boundary which encompasses all the 161 known sites in which the taxon is present, and the area of occupancy for each 162 population as the area within the extent of occurrence that is occupied by the 163 individuals. To measure the extent of occurrence, we used orthophotos 164 (obtained from REDIAM 2008 for Spain), maps (GPS GARMIN, for Morocco) 165 and AUTOCAD software (v.2006). The area of occupancy was measured by 166 multiplying the population longitudinal ranges by the average distance between 167 9 the highest and lowest heights where individuals occurred (measured at the 168 level of each sampling plot; see sampling design below). 169 170 Population size and structure 171 Population size was estimated by mutliplying the area of occupancy by plant 172 density. Density was obtained during May and June 2008 using a variable 173 number of homogeneously distributed sampling plots (2×2 m), in which we 174 recorded the number of individuals. In order to adapt the sampling strategy to 175 the fragmented and clumped nature of cliff-dwelling plants and to their 176 population sizes, the number of plots (6–30, mean 21.5; Table 1) and distances 177 between plots varied. Sampling plots were placed along the cliffs at low (0–2 m 178 from the cliff base), medium (2–4 m) and high (>4 m) transects. This plot 179 sampling system was able to cover most of the spatial range of plant 180 occurrence because either the cliffs were not too high or the individuals did not 181 occur on the higher parts. When plants were inaccessible, we used binoculars 182 to record the number of individuals. When populations occurred in a very 183 narrow valley and plants grew on both wall sides (KAN and LAUF), plots were 184 established in two parallel transects on each cliff. In order to test the accuracy 185 of the plot-based method for estimating population size, all of the individuals in 186 the three Spanish populations were also directly counted. 187 We also used the sampling plots to estimate population structure (proportion of 188 plants in different life history stages) as most recorded individuals were 189 sufficiently accessible to be classified into three categories: seedlings (less than 190 1 year old, usually with cotyledons), juveniles (one or, rarely, two herbaceous 191 16 rate of decline over the study period of 9.6% (λS = 0.9042; 95% CI: 0.9041–332 0.9043). This negative trend is similar to the one based on counts, and it was 333 only in part due to the contraction of the area of occupancy. The population 334 viability analysis (PVA) projected a sharp decline of the population, with >99% 335 probability of quasi-extinction (defined as N<10 individuals) during the next 100 336 years. Should the present conditions persist, the population is projected to be 337 reduced to only 10% of its current size (<30 individuals) in only 23 years. The 338 conditional total life span of the largest class (average age at death for 339 seedlings reaching that size class) was estimated to be 22 years. Therefore, S. 340 pustulatus can be considered a long-lived plant. 341 Lambda was negatively correlated with monthly mean maximum temperature 342 during the growing period (summer months excluded; r = -0.901, p = 0.037), 343 and positively correlated with number of days with rainfall or light precipitation 344 during the spring months (r = 0.896, p = 0.040). Lambda was not significantly 345 correlated with any of the other metereological variables (p > 0.05). 346 347 Discussion 348 Our study compared demographic variables related to the rarity and 349 vulnerability of a whole clade of three narrow endemic species of Sonchus 350 section Pustulati in two continents. This clade occurred across a continuous 351 area in the late Miocene-early Pliocene that is now separated into two regions 352 by the Alboran Sea (Silva 2014). 353 17 The information gathered by this study serves to update the conservation status 354 of the three taxa according to the categories and criteria of the IUCN Red List 355 (IUCN 2012). The Spanish S. pustulatus moves from “critically endangered” 356 (CR) to “endangered” (EN) but, based on the sharp decline projected for the 357 CAB population, the species is predicted to revert to the CR category by the 358 end of the century (see Appendix A: Conservation status). The remaining taxa 359 are considered as “deficient data” until we have more information on changes in 360 the extent of occurrence, area of distribution, size and/or number of populations 361 (see Appendix A: Conservation status and Table 4). 362 Beyond their IUCN conservation status, we found that the demographic 363 structure of most populations was similar, with adult plants being dominant and 364 a low proportion of seedlings and juveniles. This is a typical demographic 365 structure in rocky habitats (García et al. 2002; Larson et al. 2005), and is not 366 necessarily evidence of “population senescence”. A combination of factors such 367 as low seed germination and low seedling establishment in the few appropriate 368 microsites available on cliffs, along with early reproductive onset (Silva 2014) 369 and the longevity of individual plants (over 20 years for the largest individuals in 370 the case of one Spanish population) may explain such a pattern. 371 Nevertheless, two of the three Spanish populations of S. pustulatus are in 372 decline. The few existing studies on population dynamics of cliff-dwelling 373 species suggest an unusual demographic stability and resilience of cliff plants 374 (Pico & Riba 2002; García 2003) due to the much higher importance of the 375 survival of existing individuals than recruitment of new ones for the maintenance 376 of populations (Larson et al. 2000; García et al. 2008). High survival rates 377 18 translate into high individual longevity. However, our estimation of longevity in 378 S. pustulatus is much lower than in other cliff-dwelling plants (45–324 years; 379 Ehrlén & Lehtilä 2002, Forbis & Doak 2004; García, Picó, & Ehrlén 2008). This 380 fact could be species-specific, but lifespan estimation from survival rates is also 381 affected by the performance of the particular population. Lower lifespan may 382 therefore be related to the lower performance of plants in a peripheral part of 383 the geographical distribution of species of Sonchus section Pustulati, as found 384 for Silene ciliata (Giménez-Benavides, Albert, Iriondo & Escudero 2011). In any 385 case, the cliff-dwelling Sonchus studied, at least in the Baetic range, might be 386 less resilient to global changes than generally assumed for plants of very stable 387 habitats such as cliffs. 388 Due to the short distance between CAB and the other two populations, TEL and 389 ANT (1.9 and 8.2 km, respectively), similar demographic trends were expected 390 under the assumption of similar environmental factors acting in neighbouring 391 populations (Buckley et al. 2010). However, this was not entirely the case, as 392 TEL declined whereas ANT slightly increased. Differences in anthropogenic 393 disturbances at small scales and/or different relationships between population 394 dynamics and environmental factors may have promoted these divergent 395 trends. Sonchus section Pustulati species have been considered a part of the 396 pre-Mediterranean flora (Silva 2014) when a previously much wetter climatic 397 period prevailed. The ANT population is located on high cliffs protected from the 398 sun and with many points where water often oozes. Such conditions are 399 presumably more similar to the typical habitats present during species 400 diversification, suggesting that climatic refuges play a role in the persistence of 401 old floristic elements including Sonchus section Pustulati. 402 19 Because we found a strong negative influence of spring drought and mean 403 maximum temperature during the growing period on the growth rate of the 404 Spanish CAB population of S. pustulatus, it can be expected that the increase in 405 the number of days with no rainfall and drought periods predicted for the 406 southern Spanish Mediterranean (Ruiz-Sinoga, Garcia-Marin, Gabarron-407 Galeote, & Martinez-Murillo 2012) as well as the global warming will negatively 408 affect populations. 409 North Africa (Rif) represents the main distribution area for the Pustulati taxa 410 (Fig. 1) and constitutes their most plausible centre of origin and diversification 411 (Silva 2014). Although we do not know the dynamics of populations in Morocco, 412 which prevents us from comparing them with the Spanish ones, we found that 413 the latter tended to have the lowest extent of occurrence, area of occupancy, 414 and population density and size. Although both the Spanish and Moroccan 415 areas of S. pustulatus experience similar summers (dry and warm), the 416 Moroccan area is at lower latitude and is considerably wetter, having up to three 417 times higher annual rainfall than in the Spanish area (Hijmans, Cameron, Parra, 418 Jones, & Jarvis 2005). Consequently, the peripheral position of the Spanish 419 populations regarding the distribution of the Pustulati section as a whole, most 420 probably at the limit of the optimal ecological amplitude (Silva 2014), might 421 explain the demographic vulnerability of these populations to drought and high 422 temperatures. 423 Even though all populations of Pustulati section occur in a relatively small area, 424 have similar population structures, and experience similar limitations for 425 recruitment due to conditions at their specific habitats, the larger Moroccan 426 20 populations do not appear to be as threatened as the Spanish ones. The 427 evidences we present here of recent past extinctions and declining of 428 populations suggest that the Spanish populations will be more sensitive to 429 ongoing climate change. 430 431 Acknowledgments 432 Financial support for this research was provided by the Spanish Ministry of 433 Science and Innovation (grants CGL2006–00817 and CGL2010-16512). The 434 authors greatly appreciate the support in the field-work by the research group 435 led by Prof. J.F. Mota from the University of Almeria and the assistance of R. 436 Ajbilou from the University of Tetouan. We are also grateful to A.C. Brennan for 437 the linguistic revision, and one referee for valuable comments to improve a 438 preliminary version. J.L.S. was supported by a fellowship from the Spanish 439 Ministry of Science and Innovation (BES–2007–17066) and a work contract 440 from research grant CGL2010–16512. MBG was support by a national project 441 during the writing of this study (CGL2010-21642). 442 443 Appendix A. Supplementary data 444 Supplementary data associated with this article can be found, in the online 445 version, at XXXXX. 446 447 References 448 21 Boulos, L. (1973). Révision systématique du genre Sonchus L. s.l. IV. Sous-449 genre 1. Sonchus. Botaniska Notiser, 126, 155–196. 450 Buckley, Y.M., Ramula, S., Blomberg, S.P., Burns, J.H., Crone, E.E., Ehrlén, J., 451 Knigth, T.M., Pichancourt, J.-B., Quested, H., & Wardle, G.M. (2010). 452 Causes and consequences of variation in plant population growth rate: a 453 synthesis of matrix population models in a phylogenetic context. Ecology 454 letters, 13(9), 1182–1197. 455 Caswell, H. (2001). Matrix population models: construction, analysis, and 456 interpretation. Second edition. Sinauer Associates, Sunderland, 457 Massachusetts, USA. 458 Cochran, M.E., & Ellner, S. (1992). Simple methods for calculating age-based 459 life history parameters for stage-structured populations. Ecological 460 Monographs, 62, 345–364. 461 Cueto, M., Giménez E., Peñas, J., Cabello, J., Pérez-García, F.J., Jiménez-462 Sánchez, M.L., & Mota, J.F. (2003). Sonchus pustulatus. In Bañares, A., 463 Blanca, G., Güemes, J., Moreno J.C., & Ortiz S. (eds.). Atlas y libro rojo 464 de la Flora vascular amenazada de España (pp, 512–513). Madrid: 465 Dirección General de Conservación de la Naturaleza. 466 Deil, U., & De Mera, A.G. (1996). Contribution à la connaissance de la 467 phytosociologie et la biogéographie des groupements rupicoles calcaires 468 du Maroc. Bulletin de l'Institut scientifique (Rabat), 20, 87–111. 469 22 Deil, U., & Hammoumi, M. (1997). Contribution à l'étude des groupements 470 rupicoles des Bokkoya (Littoral du Rif Central, Maroc). Acta Botanica 471 Malacitana, 22, 131–146. 472 Ehrlén, J., & Lehtilä, K. (2002). How perennial are perennial plants? Oikos, 473 98(2), 308–322. 474 Fennane, M., & Ibn Tattou, M. (1998). Catalogue des plantes vasculaires rares, 475 menacées ou endémiques du Maroc. Bocconea, 8, 5–243. 476 Forbis, T.A., & Doak, D.F. (2004). Seedling establishment and life history trade-477 offs in alpine plants. American Journal of Botany, 91, 1147–1153. 478 García, M.B. (2003). Demographic viability of a relict population of the critically 479 endangered plant Borderea chouardii. Conservation Biology, 17(6), 480 1672–1680. 481 García, M.B., Guzman, D., & Goni, D. (2002). An evaluation of the status of five 482 threatened plant species in the Pyrenees. Biological Conservation, 483 103,151–161. 484 García, M.B., Picó, F.X., & Ehrlén, J. (2008). Life span correlates with 485 population dynamics in perennial herbaceous plants. American Journal of 486 Botany, 95(2), 258–262. 487 Giménez-Benavides, L., Albert, M.J., Iriondo, J.M., & Escudero, A. (2011). 488 Demographic processes of upward range contraction in long-lived 489 Mediterranean high mountain plant. Ecography, 34, 85–93. 490 23 Heywood, V.H., & Iriondo, J.M. (2003). Plant conservation: old problems, new 491 perspectives. Biological Conservation, 113(3), 321–335. 492 Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G., & Jarvis, A. (2005) Very 493 high resolution interpolated climate surfaces for global land areas. 494 International Journal of Climatology, 25, 1965–1978. 495 Horvitz, C.C., & Schemske, D.W. (1995). Spatiotemporal variation in 496 demographic transitions of a tropical understory herb: projection matrix 497 analysis. Ecological Monographs, 65(2), 155–192. 498 IUCN (2001). IUCN Red List Categories and Criteria v.3.1. IUCN Species 499 Survival Commission. IUCN, Gland, Switzerland and Cambridge, UK. 500 IUCN (2012). IUCN Red List Categories and Criteria: Version 3.1. Second 501 edition. Gland, Switzerland and Cambridge, UK: IUCN. iv + 32pp. 502 Kim, S.-C., Lee C., & Mejías, J.A. (2007). Phylogenetic analysis of chloroplast 503 DNA matK gene and ITS of nrDNA sequences reveals polyphyly of the 504 genus Sonchus and new relationships among the subtribe Sonchinae 505 (Asteraceae: Cichorieae). Molecular Phylogenetics and Evolution, 44, 506 578–597. 507 Larson, D.W., Matthes, U., & Kelly, P.E. (2005). Cliff ecology: pattern and 508 process in cliff ecosystems. Cambridge University Press. 509 Larson, D.W., Matthes, U., Gerrath, J.A., Larson, N.W.K., Gerrath, J.M., Nekola, 510 J.C., Walker, G.L., Porembski, S., & Charlton, A. (2000). Evidence for the 511 widespread occurrence of ancient forests on cliffs. Journal of 512 Biogeography, 27, 319–331. 513 24 Lavergne, S., Thompson, J.D., Garnier, E., & Debussche, M. (2004). The 514 biology and ecology of narrow endemic and widespread plants: a 515 comparative study of trait variation in 20 congeneric pairs. Oikos, 107(3), 516 505–518. 517 Morris, W., & Doak, D. (1998). Life history of the long-lived gynodioecious 518 cushion plant Silene acaulis (Caryophyllaceae), inferred from size-based 519 population projection matrices. American Journal of Botany, 85(6), 784–520 784. 521 Morris, W.F., & Doak, D.F. (2002). Quantitative conservation biology: the theory 522 and practice of population viability analysis. Sinauer Associates, 523 Sunderland, Massachusetts, USA, 480 pp. 524 Mota, J.F., Cueto, M., Pérez-García, F.J., Garrido, J.A., Martínez-Hernández, 525 F., Medina-Cazorla, J.M., Sola, A.J., & Schwarzer, H. (2005). 526 Contribución al conocimiento de la flora de Andalucía: Citas novedosas 527 de la provincia de Almería, el elemento estenócoro. Acta Botanica 528 Malacitana, 30, 227–231. 529 Picó, F.X., & Riba, M. (2002). Regional-scale demography of Ramonda myconi: 530 Remnant population dynamics in a preglacial relict species. Plant 531 Ecology, 161, 1–13. 532 REDIAM (2008). Red de Información Ambiental de Andalucía. Consejería de 533 Agricultura, Pesca y Medio Ambiente, Junta de Andalucía, Spain. 534 <http://www.juntadeandalucia.es/medioambiente/site/web/rediam> 535 [accessed in 2011]. 536 25 Ruiz-Sinoga, J.D., Garcia-Marin, R., Gabarron-Galeote, M.A., & 537 Martinez-Murillo, J.F. (2012). Analysis of dry periods along a pluviometric 538 gradient in Mediterranean southern Spain. International Journal of 539 Climatology, 32(10), 1558–1571. 540 Silva, J.L. (2014). Biology and conservation of Sonchus sect. Pustulati 541 (Asteraceae): cliff-dwelling species endemic to the Baetic-Rifan complex 542 (Western Mediterranean). Ph.D. thesis (296 pp.), Universidad de Sevilla, 543 Seville, Spain. 544 Schemske, D.W., Husband, B.C., Ruckelshaus, M.H., Goodwillie, C., Parker, 545 I.M., & Bishop, J.G. (1994). Evaluating approaches to the conservation of 546 rare and endangered plants. Ecology, 75(3), 584–606. 547 Thompson, J.D. (2005). Plant Evolution in the Mediterranean. Oxford University 548 Press, USA, 306 pp. 549 Wiens, D., Calvin, C.L., Wilson, C.A., Davern, C.I., Frank, D., & Seavey, S.R. 550 (1987). Reproductive success, spontaneous embryo abortion, and 551 genetic load in flowering plants. Oecologia, 71(4), 501–509. 552 553 Figure captions 554 Fig. 1. Overall geographical distribution of the Sonchus species (section 555 Pustulati) and areas of occupancy of populations in the Spanish Baetic (A) and 556 Moroccan Rifan ranges (B-C-D). A-B, S. pustulatus; C, S. fragilis; D, S. 557 masguindalii. Algerian range (E), S. pustulatus is unconfirmed. In the Spanish 558 area (A), the extent of occurrence is marked by a dashed line and an empty 559 0 20 40 60 80 100 Proportion of individuals (%) Populations Class-VI Class-V Class-IV Class-III Class-II Class-I Juveniles Seedlings 151 (31) 58 (52) 165 (57) 142 (13) 163 (30) 102 (46) 146 (7) 129 (2) 164 (1) 131 (20) 71 (1) 120 (2) 227 (4) 129 (20) 185 (1) 133 (7) Spanish S. pustulatus Moroccan S. pustulatus S. fragilis S. masguindalii Figure2_Black&White AppendixA Click here to download Supplementary material: AppendixA.docx AppendixA_Table3 Click here to download Supplementary material: AppendixA_Table3.docx AppendixA_Table4 Click here to download Supplementary material: AppendixA_Table4.docx AppendixA_Figure3_Caption Click here to download Supplementary material: AppendixA_Figure3_Caption.docx AppendixA_Figure3_OnlyColor Click here to download Supplementary material: AppendixA_Figure3_OnlyColor.pptx