Acta Botanica 1-2015 - za web.indd ACTA BOT. CROAT. 74 (1), 2015 43 Acta Bot. Croat. 74 (1), 43–52, 2015 CODEN: ABCRA 25 ISSN 0365-0588 eISSN 1847-8476 Variability of pollen aperture heteromorphism in annual pansies (Viola Section Melanium) SA RA MAGRINI1,*, ANNA SCOPPOLA2 1 Herbarium UTV, Tuscia University, via San Camillo de Lellis, I:01100 Viterbo, Italy 2 Department of Agriculture, Forests, Nature and Energy (DAFNE), Tuscia University, via San Camillo de Lellis, I:01100 Viterbo, Italy Abstract – Pollen heteromorphism is frequent in the section Melanium of the genus Viola, in which over 80% of the species produces pollen morphs with 3 to 6 apertures. Some au- thors have pointed out that many factors can affect the proportion of the different pollen types in perennial species, and that this proportion can change among populations. This work focuses on the study of the polymorphic pollen assemblage of three annual pansies: Viola arvensis, V. kitaibeliana, and V. hymettia, and on the assessment of its variability both within a population and within the same plant. In all the species, with both large and small fl owers, 3-, 4- and 5-aperturate pollen grains were observed, with a large prevalence of 4-aperturate types. No pollen grains with 6 apertures were found. No signifi cant vari- ability of the pollen assemblage among fl owers of the same plant was observed. In addi- tion, in these three Viola species the frequencies of the various pollen morphs are also fairly constant among plants of the same population. Keywords: annual pansies, pollen morph, pollen heteromorphism, section Melanium, variability, Viola Introduction Viola L. section Melanium Ging. is a derived, monophyletic and morphologically well- defi ned group of about 80–100 perennial and annual species showing highly-reduced ge- netic divergence (YOCKTENG et al. 2003). Its geographical distribution extends over Europe and westernmost Asia, with a few species in Northern Africa and one disjointed and proba- bly native species in North America (CLAUSEN et al. 1964, YOCKTENG et al. 2003). Pollen heteromorphism, defi ned as the production of several pollen grain morphs with different aperture numbers by the same plant (TILL-BOTTRAUD et al. 1995), occurs in over 30% of angiosperm species (MIGNOT et al. 1994). It is particularly common in this section of the genus Viola, in which over 80% of the species produce pollen morphs with 3 to 6 * Corresponding author, e-mail: magrini@unitus.it Copyright® 2015 by Acta Botanica Croatica, the Faculty of Science, University of Zagreb. All rights reserved. MAGRINI S., SCOPPOLA A. 44 ACTA BOT. CROAT. 74 (1), 2015 apertures (DAJOZ et al. 1993, TILL-BOTTRAUD et al. 1999, NADOT et al. 2000). Generally, 4- or 5-aperturate pollen grains represent more than 85% and very often more than 95% (TILL- BOTTRAUD et al. 1999). The most frequent pollen type is always 4-aperturate for perennial pansies and for typically large-fl owered annuals like Viola tricolor L., V. hymettia Boiss. & Heldr., and V. roccabrunensis Espeut (CLAPHAM et al. 1987, TILL-BOTTRAUD et al. 1999, ES- PEUT 2004), as it is related to entomophilous pollination. On the other hand, in V. arvensis Murray and V. kitaibeliana Schult. 5-aperturate pollen grains prevail (PETTET 1964, CLAPHAM et al. 1987, RANDALL 2004), which is related to autogamous pollination strategy. Still, 4-ap- erturate pollen grains were also observed by other authors (ERDTMAN 1952, DAJOZ et al. 1995, ESPEUT 1999). DAJOZ et al. (1991), MIGNOT et al. (1994) and TILL-BOTTRAUD et al. (1999) pointed out that many genetic and environmental factors can affect the proportion of the different pol- len types in heteromorphic perennial species, and that this can change among populations. On the other hand, to our knowledge, studies of the polymorphic pollen assemblage of an- nual pansies and the assessment of its variability within a population, or within the same plant, have not been previously reported in literature. Furthermore, SCOPPOLA and LATTANZI (2012), in accordance with KRISTOFFERSON (1923) and ERBEN (1985), observed an increasing reduction in the fl ower size from the lower to the upper parts of the specimen in all the analyzed species, both autogamous and heteroga- mous, and in V. arvensis and V. tricolor also from the main axis to the branches. As pollen heteromorphism is related to pollination strategy, can the main pollen type change within the same plant in such variable fl owers? The aims of this paper were: 1) to analyze the pollen assemblage of three annual pan- sies, Viola arvensis, V. kitaibeliana and V. hymettia; 2) to assess its variability among fl ow- ers within the same plant and among plants within the same population; and particularly, 3) to assess whether the main pollen type is directly correlated with the increasing reduction of the fl ower size in those species, like V. arvensis, that show both autogamous and heter- ogamous pollination, and in entomophilous species like V. hymettia. Materials and methods Study species Viola arvensis (2n = 34), the fi eld pansy, is the most common of the annual species closely related to V. tricolor, the wild pansy. It is a Mediterranean-Eurasiatic element, re- garded as an archaeophyte and widespread throughout almost the whole of Europe and SW Asia, from 0 to 1,500–1,800 m a.s.l., and as a weed linked to open shrubland and synan- thropic habitats (GAMS 1926, VALENTINE et al. 1968, MARCUSSEN and KARLSSON 2010, VOLL- RATH 2011). The corolla is shorter than or equal to the calyx (or, rarely, a little longer), usu- ally pale creamy-yellow (Fig. 1A). It is a mostly autogamous species with a very small stylar fl ap (labellum) and the entrance of the stigmatic cavity, in a front view, is obliquely forward directed (SCOPPOLA and LATTANZI 2012). Viola kitaibeliana (2n = 16), the dwarf pansy, is a Mediterranean-Caucasian species which extends to central Europe where it is a component of early stages of grassland, stony slopes and screes, found also on sandy soils, fallow land, fi elds and other open places, from 0 to approximately 1,850 m a.s.l. (WERNER 1988, RANDALL 2004, VOLLRATH 2011, MAGRINI VARIABILITY OF POLLEN HETEROMORPHISM IN ANNUAL PANSIES ACTA BOT. CROAT. 74 (1), 2015 45 and SCOPPOLA 2013). It is a small-fl owered and mostly autogamous species, with a corolla cup-shaped, not (or lightly) exceeding the calyx (Fig. 1B); the stylar fl ap is absent and the entrance of the stigmatic cavity, in a front view, is obliquely forward directed (ERBEN 1985, SCOPPOLA and LATTANZI 2012). Viola hymettia (2n = 16) is a SE-European species which is widespread mainly in Greece and the Aegean Islands, in Central and Southern Italy, where it has a disjointed dis- tribution, and in Sicily; it is an early element of stony pastures, dry open habitat and scrub fringes, occurring from 200 to 800–1,000 m a.s.l. (MERXMÜLLER 1982, ERBEN 1985, RAUS 1986, DAVIS et al. 1988). This relatively large-fl owered species has a corolla distinctly ex- ceeding its calyx, creamy and yellow colored, often suffused with violet (Fig. 1C). The stigmatic opening, in a front view, is obliquely upwards directed (SCOPPOLA and LATTANZI 2012). Collection and study sites Seedlings of V. arvensis and V. kitaibeliana were collected in wild populations in Cen- tral Italy, in February and March 2013, respectively (Tab. 1). They were transplanted in pots, together with the soil collected in situ, and cultured ex situ until May 2013. Viola hy- met tia plants were studied in situ from January to April 2013. About 100 specimens belong- ing to these taxa were used for palynological investigations. Samples of each species are deposited in Herbarium UTV, Tuscia University (Viterbo, Italy). Fig. 1. Flowers of A) Viola arvensis, B) V. kitaibeliana, and C) V. hymettia from wild populations of Central Italy. Tab. 1. Locations of the studied Viola populations. Species Localities Elevation Coordinates Habitat Viola arvensis Acque Albule, Tivoli (Roma, Italy) 66 m a.s.l. 41°57'47"N 12°42'53"E shrubby grassland in uncultivated land (calcareous soil) Viola kitaibeliana Le Vigne, Ofena (L’Aquila, Italy) 481 m a.s.l. 42°18'03"N 13°44'05"E arid and stony shrubby grassland (calcareous soil) Viola hymettia Riello, Viterbo (Viterbo, Italy) 323 m a.s.l. 42°25'38"N 12°05'38"E uncultivated escarpment with pine trees and shrubs (volcanic soil) MAGRINI S., SCOPPOLA A. 46 ACTA BOT. CROAT. 74 (1), 2015 Fig. 3. Microphotographs of the observed pollen morphs: A) 3-aperturate, B) 4-aperturate, and C) 5-aperturate grains of Viola hymettia (scale bar = 10 μm). Palynological study Pollen grains were sampled from 10–30 living plants per population that had been previously labeled and numbered. Flowers were harvested possibly before full bloom- ing to collect pollen from closed anthers, from 1 to 9 different fl owers per plant (only from the main axis) (Fig. 2). All the fl owers were dissected immediately after harvest- ing to avoid pollen loss. The plant code (species and number) and the fl ower posi- tion were registered for each one. Anthers were removed from the fl ower and put directly on a microscope slide, im- mersing them in a drop of lactic acid to let out all the pollen grains. Observations of the entire pollen assemblage were carried out under a light microscope Leitz HM- LUX3 with a magnifi cation of 100× (reso- lution: 0.25 μm) and the different pollen morphs were counted according to this scheme: • triangular shape = 3-aperturate grain, • squared shape = 4-aperturate grain, • pentagonal shape = 5-aperturate grain, • hexagonal shape = 6-aperturate grain, • round, elliptical, irregular shape = imma- ture, aborted or unidentifi able grains. This usually entailed the counting of 100–1,600 grains, or even more. For each species, microphotographs of pollen grains were taken using a digital camera, Fujifi lm FinePix S2980. Statistical analysis For statistical analysis only the sets of data with more than 150 pollen grains and with percentages of immature or aborted grains less than 20% were used. Data were analyzed by one-way ANOVA using GraphPad Prism 5.1, followed by Tukey’s multiple comparison test Fig. 2. Sample of Viola hymettia plant with the progressive numbering of the fl owers according to their arrangement from the lower to the upper parts of the stem (I–IV). The fourth fl ower is a bud suit- able for pollen collection. VARIABILITY OF POLLEN HETEROMORPHISM IN ANNUAL PANSIES ACTA BOT. CROAT. 74 (1), 2015 47 to test the signifi cant differences among fl owers in each plant and among plants in the same population. Pearson’s correlation test was performed to test for correlation between the per- centages of the main pollen morph and the position of the fl owers at 95% confi dence interval. Results Pollen heteromorphism In the examined fl owers, 3-, 4- and 5-aperturate grains were observed (Fig. 3), while no 6-aperturate grains were found. In all the species, a statistically signifi cant prevalence of the 4-aperturate grains (on average more than 93%; p < 0.0001) over the other morphs was observed. Particularly, in V. arvensis it ranges from 81% to 99% (93.58 ± 4.99%, mean ± standard deviation), with 1–18% of 5-aperturate grains (6.18 ± 5.11%) and with less than 3% of 3-aperturate ones (0.25 ± 0.57%) (Fig. 4). In V. kitaibeliana almost all the grains ob- served were 4-aperturate (98.41 ± 1.13%) while 3- and 5-aperturate grains accounted for fewer than 2% (0.47 ± 0.76% and 1.12 ± 1.13%, respectively) (Fig. 4). V. hymettia showed the greater variability with about 77–98% of 4-aperturate grains (90.74 ± 6.55%), a little greater percentage of 3-aperturate grains than the other species, up to 15.5% (1.34 ± 3.53%), and with 5-aperturate grains that range from 0 to 21% (7.92 ± 6.16%) (Fig. 4). Fig. 4. Frequencies of the different pollen morphs in all the analyzed fl owers of Viola arvensis, V. kitaibeliana, and V. hymettia. Data are ordered from the fi rst to the last fl owers. MAGRINI S., SCOPPOLA A. 48 ACTA BOT. CROAT. 74 (1), 2015 Pollen heteromorphism variability No signifi cant differences in the proportions of the various morphs were observed among fl owers of the same plant in these three species (Tabs. 2, 3, 4). Viola hymettia showed the larger variability in the polymorphic pollen assemblage, independently of fl ow- er position (Fig. 5, Tab. 4), as no trend in the variation of the percentages of the different pollen morphs was observed. On the other hand, a certain variability in the pollen assem- blage of the fi rst 2 fl owers was observed in V. arvensis, with 4-aperturate morphs ranging from about 81% to 100% and with 5-aperturate morphs from 0% to 18%. This variability was progressively and strongly reduced up to the last fl owers (Fig. 5, Tab. 2). Viola kitai- beliana developed a maximum of 3 fl owers per plant showing less variability in the pollen assemblage, with 4-aperturate morphs ranging from about 96% to 100% and 5-aperturate from 0 to 4% (Fig. 5, Tab. 3). The results of ANOVAs conducted on fl owers of different plants showed that, in V. ar- vensis, V. kitaibeliana, and V. hymettia, no signifi cant differences in the proportions of the different pollen morphs could be detected among individuals from the same population (p values vary from 0.0759 to 0.7834). In the three species, no statistically signifi cant correla- Tab. 2. Polymorphic pollen assemblages of fl owers of Viola arvensis. Mean percentage ± standard deviation and results of one-way ANOVA for three pollen morph, with statistically signifi - cance at p < 0.05 (post hoc Tukey’s test). Flower position 3-aperturate 4-aperturate 5-aperturate I 0.25 ± 0.42% 94.05 ± 4.20% 5.70 ± 4.26% II 0.16 ± 0.32% 92.77 ± 6.21% 7.07 ± 6.30% III 0.69 ± 1.32% 93.98 ± 5.11% 5.34 ± 5.58% IV 0.03 ± 0.06% 96.11 ± 2.15% 3.86 ± 2.12% V 0.11 ± 0.02% 91.73 ± 0.04% 8.16 ± 0.05% VI – 91.94 ± 0.02% 8.06 ± 0.01% p 0.4292 0.8652 0.8491 F5,48 1.001 0.3717 0.3955 R2 0.1064 0.04237 0.04496 Tab. 3. Polymorphic pollen assemblages of fl owers of Viola kitaibeliana. Mean percentage ± stan- dard deviation and results of one-way ANOVA for three pollen morph, with statistically sig- nifi cance at p < 0.05 (post hoc Tukey’s test). Flower position 3-aperturate 4-aperturate 5-aperturate I 0.29 ± 0.38% 98.13 ± 1.31% 1.59 ± 1.35% II 0.88 ± 1.09% 98.41 ± 0.98% 0.70 ± 0.80% III – 99.23 ± 1.09% 0.77 ± 1.10% p 0.3000 0.5309 0.4226 F2,13 1.361 0.6750 0.9400 R2 0.2140 0.1189 0.1582 VARIABILITY OF POLLEN HETEROMORPHISM IN ANNUAL PANSIES ACTA BOT. CROAT. 74 (1), 2015 49 tion was observed between the percentages of the main pollen morph and the position of the fl owers (Pearson’s correlation test, two-tailed p values vary from 0.1174 to 0.9362). Discussion In all three species, our results show a highly signifi cant prevalence of the 4-aperturate morph over the other two morphs. In V. hymettia the relative proportion of some pollen Fig. 5. Box-and-whisker graphs of the main pollen morphs in the fl owers of Viola arvensis, V. kitai- beliana, and V. hymettia. The ordinal numbers of the fl owers, from I to IX, are reported ac- cording to their arrangement on the main stem from the lower to the upper parts of the plant. Tab. 4. Polymorphic pollen assemblages of fl owers of Viola hymettia. Mean percentage ± standard deviation and results of one-way ANOVA for three pollen morph, with statistically signifi - cance at p < 0.05 (post hoc Tukey’s test). Flower position 3-aperturate 4-aperturate 5-aperturate I 0.43 ± 0.40% 90.42 ± 4.30% 9.15 ± 4.69% II 6.77 ± 7.64% 90.54 ± 11.27% 2.69 ± 3.71% III 0.18 ± 0.18% 89.09 ± 7.33% 10.73 ± 7.20% IV 0.28 ± 0.24% 90.54 ± 6.08% 9.18 ± 6.31% V 0.22 ± 0.28% 90.12 ± 10.44% 9.67 ± 10.68% VI 0.27 ± 0.01% 92.25 ± 0.10% 7.48 ± 0.09% VII – 88.08 ± 0.01% 11.92 ± 0.02% VIII 1.46 ± 0.07% 98.22 ± 0.02% 0.32 ± 0.02% IX 0.15 ± 0.02% 93.41 ± 0.10% 6.44 ± 0.08% p 0.0817 0.6869 0.1877 F8,19 2.174 0.7010 1.621 R2 0.4914 0.2376 0.4187 MAGRINI S., SCOPPOLA A. 50 ACTA BOT. CROAT. 74 (1), 2015 morphs varies quite considerably: in some fl owers, from sunny areas, there were no 3-aper- turate pollen grains at all, while in more pigmented fl owers and in others with upper petals suffused with violet, in the shade of the trees, more than 15% were 3-aperturate pollen grains and only 77% 4-aperturate ones. Variability in the color of the corolla related to light gradient is already known for V. tricolor, an annual pansy with creamy to bluish or violet large petals occurring in meadows and habitats linked to woodlands (PETTET 1964). TILL- BOTTRAUD et al. (1999) proved that pollen-type proportions of a perennial pansy, V. cal- carata L., vary among populations along altitudinal gradients. On the other hand, to our knowledge there is no information regarding pollen heteromorphism variability within the same population of annual pansies linked to gradient. The pollen assemblage of the two other species is more uniform than in V. hymettia. Vi- ola arvensis shows a certain variability in the percentages of the main morph, especially in the fi rst fl owers, where 4-aperturate grains range from 81% to 100%. A greater variability, ranging from 69% to 96%, was observed among those populations with mostly 5-aperturate grains (PETTET 1964). In the pollen assemblage of V. kitaibeliana almost all the grains are 4-aperturate, in accordance with a recent study from Iran (SAEIDI MEHRVARZ et al. 2014). By contrast, our recent unpublished data highlight a rather wide variability among other popu- lations from Central Italy where 5-aperturate grains prevail (from 83 to 92%). The pollen assemblages of the last two species show less variability when 4-aperturate pollen grains predominate. Our study was focused on the assessment of the pollen heteromorphism variability in these three Viola species, taking into account differences within an individual and within a population. Pollen heteromorphism is related to pollination strategy: in heterogamous plants there is a prevalence of pollen grains with a few apertures (generally 4 or less), while in the autogamous ones 5-aperturate morphs prevail (DAJOZ et al. 1993). So, we hypothe- sized a variability among fl owers related to the increasing reduction of their size from the lower to the upper parts of the plant in those species, like V. arvensis, that show both au- togamous and heterogamous pollination. We even expected to observe the main pollen type changing among such variable fl owers within the same plant. On the other hand, we did not expect to fi nd this change in species like V. hymettia showing fl ower features so adapted to entomophilous pollination (large and fl attened corolla, scent, stigmatic opening up-wards directed, protruding labellum, etc.). Our results have only confi rmed the last hypothesis. In fact, no signifi cant correlation was detected between the number of apertures and the posi- tion of the fl ower, either in V. hymettia or in two other species. No signifi cant differences were recorded even among individuals. Our results are not in accordance with data reported by DAJOZ et al. (1995) for some V. arvensis populations where the percentages of the different pollen morphs vary signifi cantly among plants of the same population so much as to change the main morph. These confl icting fi ndings are prob- ably due to differences in sampling methods that may often lead to biased estimates, espe- cially the analysis of a maximum of only 200 pollen grains per fl ower without any indica- tion about the developmental stage of the collected fl owers. There is a large difference in the number of pollen grains between fl owers in full bloom and fl ower buds (from about 60 to over 2,500). In fact, only in the closed anthers collected from buds can all the pollen as- semblage be found, although with a variable percentage of unidentifi able grains. Finally, according to our results, even when there is variability in the relative frequen- cies of the different pollen morphs, the same type always prevails within both a single plant VARIABILITY OF POLLEN HETEROMORPHISM IN ANNUAL PANSIES ACTA BOT. CROAT. 74 (1), 2015 51 and the population. Then, it is possible to identify the main pollen morph by analyzing just a few fl owers per plant or per population. However, in order to avoid bias due to sampling methods, we suggest the collection of fl ower buds in well developed plants, with at least three or more fl owers, and the analysis of the whole pollen assemblage rather than a few hundred grains. Acknowledgements Thanks are due to E. Chiummariello for assistance in plant collecting, in fl ower sam- pling and pollen grain recording, and to S. Buono and T. Kirin for images of Viola hymettia and V. arvensis fl owers, respectively. We thank also the anonymous reviewers for their con- structive comments. References CLAPHAM, A. R., TUTIN, T. G., MOORE, D. M., 1987: Flora of the British Isles. 3rd ed., 107– 112. Cambridge University Press, Cambridge. CLAUSEN, J., CHANNELL, R. B., NUR, U., 1964: Viola rafi nesquii, the only Melanium violet native to North America. Rhodora 66, 32–46. DAJOZ, I., MIGNOT, A., HOSS, C., TILL-BOTTRAUD, I., 1995: Pollen aperture heteromorphism is not due to unreduced gametophytes. American Journal of Botany 82, 104–111. DAJOZ, I., TILL-BOTTRAUD, I., GOUYON, P. H., 1991: Evolution of pollen morphology. Sci- ence 253, 66–68. DAJOZ, I., TILL-BOTTRAUD, I., GOUYON, P. H., 1993: Pollen aperture polymorphism and ga- metophyte performance in Viola diversifolia. Evolution 47, 1080–1093. DAVIS, P. H., MILL, R. R., TAN, K., 1988: Viola L. In: DAVIS, P. H., MILL, R. R., TAN, K. (eds.), Flora of Turkey and the East Aegean Islands (Suppl. 1), Vol. 10, 62–64. Edin- burgh University Press, Edinburgh. ERBEN, M., 1985: Cytotaxonomische Untersuchungen an südosteuropeischen Viola-Arten der Sektion Melanium. Mitteilungen der Botanischen Staatssammlung München 21, 339–740. ERDTMAN, G., 1952: Pollen morphology and plant taxonomy. Almqvist & Wiksell, Stock- holm. ESPEUT, M., 1999: Errata et addenda de l’article: »Approche du genre Viola dans le Midi Meditérranéen français«. Le Monde des Plantes 467, 7–9. ESPEUT, M., 2004: Viola roccabrunensis sp. nov. Le Monde des Plantes 482, 18–21. GAMS, H., 1926: Viola L. In: HEGI, G., Illustrierte Flora von Mitteleuropa, vol. 5, 586–656. Lehmanns Verlag, München. KRISTOFFERSON, K. B., 1923: Crossings in Melanium-violets. Hereditas 4, 251–289. MAGRINI, S., SCOPPOLA, A., 2013: Is it possible to defi ne the real Italian distribution area of the annual pansy, Viola kitaibeliana? In: PECCENINI, S., DOMINA, G. (eds.), Contributi alla ricerca fl oristica in Italia, 53–54. Società Botanica Italiana, Palermo. MAGRINI S., SCOPPOLA A. 52 ACTA BOT. CROAT. 74 (1), 2015 MARCUSSEN, T., KARLSSON, T., 2010: Violaceae. In: JONSELL, B., KARLSSON, T. (eds.), Flora Nordica, vol. 6, 12–52. Royal Swedish Academy of Sciences, Stockholm. MERXMÜLLER, H., 1982: Viola L. In: PIGNATTI, S. (ed.), Flora d’Italia, Vol. 2, 102–118. Edag- ricole, Bologna. MIGNOT, A., HOSS, C., DAJOZ, I., LEURET, C., HENRY, J. P., DREUILLAUX, J. M., HEBERLE-BORS, E., TILL-BOTTRAUD, I., 1994: Pollen aperture polymorphism in the angiosperms: impor- tance, possible causes and consequences. Acta Botanica Gallica 141, 109–122. NADOT, S., BALLARD, H. E. JR, CREACH, J. B., DAJOZ, I., 2000: The evolution of pollen het- eromorphism in Viola: A phylogenetic approach. Plant Systematics and Evolution 223, 155–171. PETTET, A., 1964: Studies on British pansies. II. The status of some intermediates between Viola tricolor L. and V. arvensis Murr. Watsonia 6, 51–69. RANDALL, R. E., 2004: Biological fl ora of the British Isles no. 233: Viola kitaibeliana Schult(es). Journal of Ecology 92, 361–369. RAUS, T. H., 1986: Viola L. In: STRID, A. (ed.), Mountain fl ora of Greece, Vol. I, 608–640. Cambridge University Press, Cambridge. SAEIDI MEHRVARZ S., YOUSEFI N., MOHAMMADI M., MARCUSSEN T., 2014: Pollen studies in the genus Viola (Violaceae) from Iran. Acta Botanica Croatica 73, 93–106. SCOPPOLA, A., LATTANZI, E., 2012: Viola section Melanium (Violaceae) in Italy. New data on morphology of Viola tricolor-Group. Webbia 67, 47–64. TILL-BOTTRAUD, I., MIGNOT, A., DE PAEPE, R., DAJOZ, I., 1995: Pollen heteromorphism in Nicotiana tabacum (Solanaceae). American Journal of Botany 82, 1040–1048. TILL-BOTTRAUD, I., VINCENT, M., DAJOZ, I., MIGNOT, A., 1999: Pollen aperture heteromor- phism. Variation in pollen-type proportions along altitudinal transects in Viola calcara- ta. Comptes Rendus de l’Académie des Sciences – Series III – Sciences de la Vie 322, 579–589. VALENTINE, D. H., MERXMÜLLER, H., SCHMIDT, A., 1968: Viola L. In: TUTIN, T. G., HEYWOOD, V. H., BURGES, N. A., MOORE, D. M., VALENTINE, D. H., WALTERS, S. M., WEBB, D. A. (eds.), Flora Europaea, vol. 2, 270–282. Cambridge University Press, Cambridge. VOLLRATH, K. H., 2011: Viola in Nordostbayern. Retrieved May 3, 2013 from http://www. regnitzfl ora.de/Veilchen_20_12_10.pdf WERNER, K., 1988: Viola kitaibeliana Schultes auf der Schwellenburg bei Erfurt – ein Neu- fund für die DDR. Hercynia 25, 142–143. YOCKTENG, R., BALLARD, H. E. JR, MANSION, G., DAJOZ, I., NADOT, S., 2003: Relationships among pansies (Viola section Melanium) investigated using ITS and ISSR markers. Plant Systematics and Evolution 241, 153–170.