Comparative Floral and Vegetative Differentiation Between Two
Total Page:16
File Type:pdf, Size:1020Kb
Pl. Syst. Evol. 262: 209–224 (2006) DOI 10.1007/s00606-006-0473-2 Comparative floral and vegetative differentiation between two European Aquilegia taxa along a narrow contact zone M. Medrano, M. C. Castellanos, and C. M. Herrera Estacio´n Biolo´gica de Don˜ ana, Consejo Superior de Investigaciones Cientı´ficas, Sevilla, Spain Received March 13, 2006; accepted July 19, 2006 Published online: November 20, 2006 Ó Springer-Verlag 2006 Abstract. As a first step in determining the identity currently established view of the radiation process and relative importance of the evolutionary forces in the genus Aquilegia in North America, where the promoting the speciation process in two closely differentiation of floral traits seems to have played related European taxa of Aquilegia, we investigated a more important role. the levels of morphological variation in floral and vegetative characters over the narrow region where Key words: Aquilegia pyrenaica subsp. cazorlensis, their ranges enter into contact, and evaluate the Aquilegia vulgaris subsp. vulgaris, floral character, relative importance of both types of traits in their Iberian Peninsula, inter- and intraspecific variation, differentiation. A total of 12 floral and ten vegeta- morphological differentiation, Ranunculaceae, tive characters were measured on 375 plants vegetative character. belonging to seven A. vulgaris populations and six A. pyrenaica subsp. cazorlensis populations located in southeastern Spain. Floral and vegetative mor- The radiation of angiosperms has been tightly phological differentiation occur between taxa and linked to the diversification in floral morphol- among populations within taxa, but only vegetative ogy and function, and this effect is particularly characters (particularly plant height and leaf pet- pronounced in animal-pollinated lineages iolule length) contribute significantly to the dis- (Grimaldi 1999, Fenster et al. 2004, but see crimination between taxa. Differentiation among Waser 1998, Gorelick 2001). Specifically, floral populations within taxa is mostly explained by characters contribute to species separation variation in floral traits. Consequently, morpho- much more often among animal-pollinated logical divergence between the two taxa cannot be taxa than among wind- and water-pollinated interpreted as an extension of among-population taxa (Grant 1949). Progress in our understand- differences occurring within taxa. Multivariate ing of floral diversification has been mainly vegetative, but not floral, similarity between pop- based on examination of phylogenetic and ulations could be predicted from geographical distance. Moreover, the key role of certain vegeta- ecological correlates of floral variation at the tive traits in the differentiation of A. vulgaris and species level and above (e.g. Barrett and A. p. cazorlensis could possibly be attributable to Graham 1997, Hapeman and Inoue 1997, the contrasting habitat requirements and stress Hodges 1997, von Hagen and Kadereit 2003, tolerance strategies of the two taxa. These pre- Graham and Barrett 2004) and studies of liminary findings seem to disagree with the phenotypic selection at the within-population 210 M. Medrano et al.: Floral vs. vegetative differentiation in two European Aquilegia level based on either naturally-occurring (e.g. promoting their differentiation. North Amer- Campbell et al. 1991, Herrera 1993, Go´mez ican species of Aquilegia (Ranunculaceae) 2000, Maad 2000) or artificially induced (e.g. have become the textbook example of adap- Herrera 2001, Aigner 2004, Castellanos et al. tive radiation driven by specialization on 2004) variation in floral traits. In contrast, different pollinators in combination with eco- there have been relatively few investigations logical niche diversification (Schluter 2000 focusing on the connections between micro- and references therein). In Europe, the genus and macroevolutionary patterns of floral var- Aquilegia comprises roughly the same number iation (Miller 1981, Barrett 1995, Johnson of species as in North America (Munz 1946, 1997), and on intraspecific floral variation Jalas et al. 1999), yet the factors underlying and its relationship to evolutionary divergence their diversification remain essentially unex- in pollinators (see however e.g. Herrera 1990, plored to date. This study was conducted Robertson and Wyatt 1990, Arroyo and Dafni along the single southeastern mountain of the 1995, Boyd 2002, Herrera et al. 2002, Silva- Iberian Peninsula where populations of Montellano and Eguiarte 2003). the widely distributed A. vulgaris coexist with Factors other than pollinators may also the narrow endemic A. p. cazorlensis. In this promote floral divergence (see Wilson and region, the two taxa are pollinated by the Thomson 1996 and references therein) or same few bumblebee species (see below), account for population differences in floral though they grow in locally different habitats. characteristics along the geographical range of In particular, given that A. p. cazorlensis a species (e.g. Galen 1999 and references occurs in more stressful sites (such as rock therein). Variation in abiotic factors, e.g. soil crevices and sandy soils around cliff bases) nutrient availability or water stress, may ulti- than A. vulgaris, we expect that the role of mately explain population differences in the certain vegetative traits (particularly those floral characteristics of some species, either related to plant or leaf size) should be directly by selection on certain floral charac- particularly important in their differentiation, teristics (Frazee and Marquis 1994, Galen et al. as found for other pairs of narrow endemic 1999, Galen 2000) or indirectly through their and widespread closely related taxa (Lavergne influence on vegetative characters allometrical- et al. 2003, 2004). Specifically, the following ly or pleiotropically linked to floral traits (Bond questions will be addressed in this paper: (1) and Midgley 1988; Midgley and Bond 1989; Do the two taxa studied exhibit discernible Andersson 1993, 1997). Studying geographical geographic variation in floral and vegetative variation on floral characters in conjunction traits at the relatively small spatial scale of with vegetative traits should provide a more this investigation? If they do, (2) which type reliable evaluation of the relative importance of of traits, floral or vegetative, are those most pollinators in the divergence process that could intraspecifically variable? (3) Do floral and be taking place within a species. vegetative traits vary in unison across popu- The main goal of this paper is to compare lations of the same species? (4) Can interspe- the levels of phenotypic morphological vari- cific differences in the traits examined be ation in floral and vegetative characters interpreted as extensions of among-popula- among populations of two closely-related tion differences occurring within species? And European taxa of the genus Aquilegia, A. (5) does the observed variation conform to vulgaris subsp. vulgaris (A. vulgaris hereafter) any discernible geographical pattern? and A. pyrenaica subsp. cazorlensis (A. p. cazorlensis hereafter), throughout a region Materials and methods where their ranges enter into close contact, as a first step in determining the identity and Study species and sites. Aquilegia vulgaris L. is a relative importance of the evolutionary forces widely distributed perennial herb whose natural M. Medrano et al.: Floral vs. vegetative differentiation in two European Aquilegia 211 populations are common throughout Eurasian flowers are bisexual, self-compatible and to some mountain forests, although it sometimes occurs in extent self-pollinating in the absence of pollinators open woodlands and meadows at or around sea (unpublished data). In preliminary pollinator cen- level. In our study sites plants grow along stream suses and observations, the predominant visitors to margins or poorly drained open meadows around both Aquilegia species at the study region are springs at 900 to 1700 m of elevation. Flowering bumblebees (Table 1), though a variety of other takes place mainly from May to early June. insects visit the flowers occasionally. Aquilegia pyrenaica subsp. cazorlensis (Heywood) This study was conducted during May-June Galiano and Rivas-Martı´nez is a narrowly endemic 2004 on seven A. vulgaris populations and six A. p. perennial herb, restricted to a few populations in cazorlensis populations located in the Natural Park the Sierras de Cazorla and El Pozo, in the Spanish Sierras de Cazorla-Segura-Las Villas, in southeast- province of Jae´n (Fig. 1), occurring from 1200 to ern Spain, where both species coexist (Fig. 1). The 1950 m of altitude. Plants grow in rifts of limestone 13 study populations were regularly scattered over outcrops and on sandy soils in shady, damp sites at the known distribution of each taxon in the region cliff bases, and bloom during June-early July. (Fig. 1) and were located along a latitudinal In both species, a mature plant consists of a gradient. While the region has a Mediterranean- slender rhizomatous stem with one to several basal type climate characterized by a marked summer rosettes, each with 3–6 pubescent ternate com- drought, with 90% of the annual precipitation pound leaves. Mature plants can produce one to falling during October-April, within the Natural several paniculate inflorescences, each bearing 1–13 Park a significant negative correlation between (A. vulgaris) or 1–8 (A. p. cazorlensis) flowers, latitude and yearly total rainfall exists (Herrera et although not all mature individuals flower