Grana ISSN: 0017-3134 (Print) 1651-2049 (Online) Journal homepage: https://www.tandfonline.com/loi/sgra20 Pollen morphology of Dioscorea (Dioscoreaceae) from the Atlantic Forest in southeast Brazil (São Paulo) with a contribution to the systematics of Neotropical species Cynthia Fernandes Pinto Da Luz, Lia de Oliveira Barbuglio-Santiago, Ana Rita G. Simões, Jonathan Henrique Da Silva, Valéria Leobina Dos Santos & Mizue Kirizawa To cite this article: Cynthia Fernandes Pinto Da Luz, Lia de Oliveira Barbuglio-Santiago, Ana Rita G. Simões, Jonathan Henrique Da Silva, Valéria Leobina Dos Santos & Mizue Kirizawa (2020): Pollen morphology of Dioscorea (Dioscoreaceae) from the Atlantic Forest in southeast Brazil (São Paulo) with a contribution to the systematics of Neotropical species, Grana, DOI: 10.1080/00173134.2020.1737730 To link to this article: https://doi.org/10.1080/00173134.2020.1737730 Published online: 13 May 2020. Submit your article to this journal View related articles View Crossmark data Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=sgra20 Grana, 2020 https://doi.org/10.1080/00173134.2020.1737730 Pollen morphology of Dioscorea (Dioscoreaceae) from the Atlantic Forest in southeast Brazil (São Paulo) with a contribution to the systematics of Neotropical species CYNTHIA FERNANDES PINTO DA LUZ 1, LIA DE OLIVEIRA BARBUGLIO-SANTIAGO1, ANA RITA G. SIMÕES 2,3, JONATHAN HENRIQUE DA SILVA3, VALÉRIA LEOBINA DOS SANTOS1 & MIZUE KIRIZAWA3 1 Núcleo de Pesquisa em Palinologia, Instituto de Botânica, São Paulo, Brazil, 2Systematic and Evolutionary Botany Lab, Ghent University, Gent, Belgium, 3Núcleo de Pesquisa Curadoria do Herbário SP, Instituto de Botânica, São Paulo, Brazil Abstract Dioscorea is a highly diverse genus in Brazil, particularly the south-eastern region, and most especially in the Atlantic Forest. Recent phylogenetic studies have helped clarify the intricate subgeneric classification of the genus, although some species remain to be sampled, and further studies are necessary to ascertain their systematic placement in the genus. Here, a palynological description of nine species occurring in a patch of Atlantic Forest, in the State of São Paulo, is provided. Three of these have not yet been sampled in molecular studies. The pollen of D. tauriglossum had never been characterised before. Dioscorea pollen are spheroidal to ellipsoidal, small to medium in size (the larger dimensions in D. dodecaneura and the smaller in D. amaranthoides), with disulculate aperture. The pollen wall ornamentation contributed to the segregation of two species – psilate-perforate (D. altissima, syn. D. choncrocarpa) and rugulate-perforate (D. monadelpha), as well as the separation of the other seven species in two groups – microreticulate (D. dodecaneura, D. multiflora, D. tauriglossum), and striate (D. amaranthoides, D. laxiflora, D. olfersiana, D. trilinguis). Systematic implications of the palynological variation found are discussed under the light of recent molecular phylogenetic studies. Keywords: Eudioscorea subgenera, Helmia subgenera, monocotyledons, pollen characters, yams The current circumscription of Dioscoreaceae R. Br., based on phylogenetic analyses, includes four genera which form a strongly supported monophyletic group: Dioscorea L., Stenomeris Planch., Tacca J.R. et G. Forst. and Trichopus Gaertn. (including Avetra H. Perrier) (Caddick et al. 2000, 2002). Palynology has shown to be important for improving the morphological characterisation of these genera (Caddick et al. 1998). The palynological characters which define this group include a glandular tapetum, simultaneous microsporogenesis (undocumented in Trichopus zeylanicus) and differences in the exine ornamentation and aperture number: the pollen grains of Dioscorea are monosulcate or disulcate, perforate-striate, perforate-finely reticulate or perforatereticulate and occasionally gemmate; the pollen grain of Tacca is monosulcate and finely reticulate; the pollen grain of Stenomeris is monosulcate, perforate to finely reticulate; the pollen grain of Trichopus is monosulcate and spinulate/microechinate, and the pollen grain of Avetra is 4- or 5-pantoporate and spinulate/microechinate (Caddick et al. 1998). Dioscorea (including Borderea Miegev., Epipetrum Phil., Nanarepenta Matuda, Rajania L. and Tamus L.), holds 95% of the taxonomic diversity of the family, with c. 350–400 herbaceous species and lianas, and is found particularly in the tropics and subtropics (Caddick et al. 2002; Govaerts et al. 2007; Kirizawa & Xifreda 2009; Kirizawa et al. 2015). It is a highly diverse genus in Brazil, with 140 species present, of which 104 are endemic. The greatest diversity is concentrated in the south-eastern region, with 97 species present, 69% of the Brazilian taxa, (followed by the central-west region with 55 species) and in the Atlantic Forest, where 81 Correspondence: Cynthia Fernandes Pinto da Luz, Núcleo de Pesquisa em Palinologia, Instituto de Botânica, São Paulo, Brazil. E-mail: cyluz@yahoo.com.br (Received 25 August 2019; accepted 14 February 2020) © 2020 Collegium Palynologicum Scandinavicum 2 C. F. P. da Luz et al. species occur, concentrating 58% of the Brazilian species (followed by Cerrado, with 68 species) (Kirizawa et al. 2015). Therefore, the southeast region, and the Atlantic Forest biome, in particular, contribute in a very significant way to the overall diversity of the genus in Brazil. In the State of São Paulo, 49 species of Dioscorea have been recorded, occurring mostly in the Dense Ombrophilous Montane Forest in the eastern region of the State (Kirizawa & Xifreda 2011; Kirizawa et al. 2015). Knuth (1924) proposed a subdivision of Dioscorea in 60 sections, a great part of which presented delimitation problems due to morphological similarities. Other factors have contributed to the intricacies in the taxonomy at species level in this genus, such as the difficulty in sampling climbing species, especially the dioecious ones, and the description as distinct species of plants with both pistillate and staminate flowers. Recent molecular phylogenetic and biogeographical studies have helped clarify the intricate subgeneric classification of the genus, previously based primarily on morphological characters, by suggesting novel hypotheses of species relationships (Viruel et al. 2016; Couto et al. 2018). While the work of Viruel et al. (2016) outlined major clades across Dioscorea, Couto’s study provided an expanded sampling of Neotropical species that improved greatly our understanding of the systematics of the Brazilian species. However, species have remained to be sampled, and further studies are necessary to ascertain their systematic placement in the genus. Schols et al. (2001) have emphasised the importance of pollen characters to investigate the relationships within Dioscorea. The palynological research in Dioscoreaceae is long dated and several studies have described the pollen of Dioscorea, such as Erdtman (1952), Ikuse (1956), Miège (1965), Sharma (1967), Huang (1970), Heusser (1971), Barroso et al. (1974), Zavada (1983), Gupta and Sharma (1986), Su (1987), Roubik and Moreno (1991), Corrêa (1996), Caddick et al. (1998), Xifreda (2000), Schols et al. (2001, 2003, 2005a, 2005b), Wilkin et al. (2002, 2009), Couto et al. (2014), Alzer (2017) and Sheikh and Kumar (2018). With the purpose of contributing to the characterisation of the Brazilian species of Dioscorea, we have investigated the importance of pollen morphology in the delimitation of species and/or species groups in the genus. A palynological description of nine species occurring in a patch of Atlantic Forest, in the State of São Paulo, is provided – D. altissima Lam (syn. D. chondrocarpa), D. amaranthoides C. Presl., D. dodecaneura Vell., D. laxiflora Mart. ex. Griseb., D. monadelpha (Kunth) Griseb., D. multiflora Mart. ex. Griseb., D. olfersiana Klotzsch ex Griseb., D. tauriglossum R. Knuth, and D. trilinguis Griseb. – belonging in subgenera Helmia (Kunth) Benth. and Eudioscorea Pax (Silva et al. in press), and distributed across seven sections as recognised by Knuth (1924). The shape of the fruits and the disposition of the wings on the seeds, along with other vegetative and reproductive characters, have been essential for placing the studied species in these two subgenera, and for the inclusion of such taxa in the corresponding sections by Knuth (1924). Considering the molecular phylogeny of Couto et al. (2018), these nine species are distributed across three distinct clades (see Table I). Five of these species are endemic to Brazil: Dioscorea laxiflora, D. monadelpha, D. olfersiana, D. tauriglossum and D. trilinguis. Three of these species have not yet been sampled in molecular studies – D. altissima (syn. D. chondrocarpa), D. tauriglossum and D. trilinguis; they were missing in the studies of Viruel et al. (2016), Couto et al. (2018), and no GenBank accessions were located either. Among these, D. tauriglossum was, for the first time, palynologically analysed. Table I. List of the species included in the sampling of the current palynological study, with additional information on their systematics: subgenera and sections proposed by Knuth (1924), and clades resolved in Couto et al. (2018). Species are listed according to their molecular systematic placement and exine type, to help the interpretation of the results; species endemic to Brazil are highlighted in bold; the species analysed for the first time is marked with an asterisk (*). Species Dioscorea amaranthoides Dioscorea laxiflora Dioscorea olfersiana Dioscorea trilinguis Dioscorea multiflora Dioscorea monadelpha Dioscorea dodecaneura *Dioscorea tauriglossum Dioscorea altissima (syn. D. chondrocarpa) Exine Striate in two directions Striate in two directions Striate in two directions Striate in one direction Microreticulate Rugulate-perforate Microreticulatee Microreticulate Psilate-perforate Clade (Couto et al. 2018) Subgenus Section New World I New World I New World I Not sampled New World I New World II New World III Not sampled Not sampled Dioscorea Dioscorea Dioscorea Dioscorea Helmia Helmia Dioscorea Helmia Helmia Strutantha Uline Cryptantha Uline Cryptantha Uline Periandrium Sphaerantha Uline Monadelpha Uline Lasiogyne Dematostemon Chondrocarpa Pollen morphology of Dioscorea The taxonomic sampling in the present work is phylogenetically diverse, since the six species already sampled in previous molecular studies belong in three distinct clades as recognised by Couto et al. (2018): New World I, New World II, and New World III (Table I). Overall, the palynological characterisation here provided offers novel data for one species, as well as an important contribution of pollen characters to the systematics of Dioscorea, from an area that is within the most diverse region of Brazil for the genus. Material and methods The sampling in the present work results from general collections undertaken during a floristic survey of a large area of Atlantic Forest surrounding the city of São Paulo, preceding the construction of a highway in that particular region. Plants were collected along an area that is now crossed by the southern and northern stretches of the Rodoanel Mario Covas (highway SP-21), São Paulo, Brazil. The construction of highway SP-21 started in 1998 and is one of the largest roads entreprises of the Government of the State of São Paulo. It contains four stretches, planned to alleviate heavy traffic, and connects ten different roads. During this construction, the researchers of Instituto de Botânica (São Paulo) worked ahead of the road developments, to rescue and document all the species that refer to the Atlantic Forest, resulting in a complete survey of the region’s vegetation. The specimens of Dioscorea which were used in the present study, were collected in the municipalities of Arujá, Guarulhos, Itapecerica da Serra, Mauá, São Bernardo de Campo and São Paulo, São Paulo State, Brazil, from 2008 to 2015. Nine species were surveyed, represented by 27 specimens (listed under ‘Specimens investigated’). Collected samples have been labelled by abbreviation of the species’ epithet and addition of the last two numbers of the herbarium voucher. Specimens from other locations were also consulted for data comparison. Herbarium acronyms of analysed specimens follow the Index herbariorum (which is continuously updated). The flower bud samples were removed from specimens from the Herbario Científico do Estado ‘Maria Eneyda P. Kauffmann Fidalgo’ (SP) of the Instituto de Botânica. Pollen grains were acetolysed following Erdtman (1960) and their dimension average values were derived from measurements of 25 pollen grains per sample. Measurements of others morphological structures (sulcus length, sexine and nexine thickness) were based upon ten pollen grains per sample. 3 In all cases, three specimens per taxon were analysed, to accommodate intraspecific variation. Statistical analyses conducted include arithmetic average (x), average standard deviation (sx), sample standard deviation (s), coefficient of variability (V%) and 95% confidence interval (CI) (Vieira 2008). To test whether these pollen grain characteristics allow for the grouping of species, a principal component analysis (PCA) was performed using the software FITOPAC (Shepherd 1996) and PC-ORD version 7 (McCune & Melfford 1999). For this analysis, we used seven metric variables: larger equatorial axis in polar view (VPEM), shorter equatorial axis in polar view (VPEm), equatorial axis in equatorial view (VEEM), polar axis in equatorial view (VEEm), sulcus length (SUCC), sexine thickness (SEX) and nexine thickness (NEX). To understand how Dioscorea specimens relate to each other by pollen morphology, a cluster analysis (CA) was performed using the software PC-ORD version 7 (McCune & Melfford 1999). This produced a similarity dendogram by calculating the Euclidean distance, based on the same seven PCA metric variables. Pollen grains were analysed under light microscopy (LM) and photomicrographs were produced using an Olympus BX 50 microscope with a coupled video camera, and assistance of Olympus software CellSens for Windows 1.5. In addition, images of the acetolysed pollen grains were obtained using a Philips XL 20 scanning electron microscope at the Instituto de Botânica and a Quanta FEG 250 FEI scanning electron microscope at the Universidade Federal de São Paulo (São Paulo). Acetolysed pollen grains were dehydrated by spreading ethanol directly on the microscope stubs, which were then sputter-coated with a 20 nm thick gold layer. Terminology follows Punt et al. (2007) and Hesse et al. (2009). Microscope slides with the pollen material are deposited at the palynotheca of the ‘Núcleo de Pesquisa em Palinologia’ of the Instituto de Botânica. Results General description The Dioscorea pollen grains varied from small to medium sized (longest axis reaching 16.7–36.6 µm) (Tables II, III), spheroidal to ellipsoidal in polar view, slightly biconvex or flattened/convex in equatorial view (Figures 1–5). Pollen is disulculate, with two elongated apertures parallel to the longest axis of the grain, situated between the proximal and distal poles. The sulculi may be slightly displaced from the equator, in which case one of the poles is larger than the other, being subisopolar. The sexine is 4 C. F. P. da Luz et al. Table II. Measures (in micrometres) in polar view of Dioscorea L. pollen grains through light microscopy (n = 25). Larger equatorial axis in polar view (VPEM) Species Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea altissima (SP 304317) altissima (SP 446562) altissima (SP 487481) amaranthoides (SP 49534) amaranthoides (SP 168574) amaranthoides (SP 446580) dodecaneura (SP 298281) dodecaneura (SP 246876) dodecaneura (SP 276042) laxiflora (SP 202296) laxiflora (SP SP 248709) laxiflora (SP 287969) monadelpha (SP 426529) monadelpha (SP 472164) monadelpha (SP 487480) multiflora (SP 289468) multiflora (SP 444406) multiflora (SP 475315) olfersiana (SP 318370) olfersiana (SP 377050) olfersiana (SP 474884) tauriglossum (SP 248719) tauriglossum (SP 472163) tauriglossum (SP 472166) trilinguis (SP 53704) trilinguis (SP 212491) trilinguis (SP 426533) Shorter equatorial axis in polar view (VPEm) x sx s V% CI x sx s V% CI 30.9 29.4 29.6 22.0 20.6 16.7 35.4 36.6 34.1 29.2 27.1 29.5 30.9 28.5 29.4 32.2 29.3 28.9 28.6 26.0 24.5 26.9 25.5 27.5 28.5 32.0 31.9 0.7 0.4 0.5 0.4 0.3 0.5 0.5 0.6 0.5 0.2 0.3 0.3 0.4 0.3 0.4 0.6 0.3 0.3 0.3 0.3 0.3 0.4 0.4 0.3 0.2 0.2 0.1 3.5 1.8 2.3 1.8 1.5 2.4 2.7 2.9 2.5 1.2 1.4 1.6 2.0 1.6 2.0 2.9 1.7 1.6 1.7 1.4 1.4 1.9 1.8 1.6 1.1 1.1 0.7 11.2 6.1 7.8 8.2 7.1 14.5 7.5 8.0 7.3 4.2 5.0 5.4 6.4 5.7 7.0 9.1 5.9 5.7 6.0 5.5 5.7 7.1 7.1 5.7 3.8 3.4 2.3 29.5–32.4 28.7–30.2 28.7–30.6 21.3–22.8 20.0–21.2 15.7–17.7 34.3–36.5 35.4–37.9 33.0–35.1 28.7–29.7 26.6–27.7 28.8–30.1 30.1–31.7 27.8–29.2 28.6–30.3 30.9–33.4 28.6–30.0 28.2–29.5 27.9–29.3 25.4–26.6 23.9–25.1 26.1–27.7 24.8–26.3 26.8–28.1 28.1–29.0 31.6–32.5 31.6–32.2 23.7 23.8 23.9 17.3 14.8 11.8 24.7 21.1 27.1 18.9 21.0 21.7 26.2 21.5 21.4 22.3 24.6 21.2 20.7 21.1 19.4 21.0 21.0 22.2 21.3 26.8 26.2 0.5 0.5 0.5 0.4 0.3 0.3 0.7 0.7 0.6 0.6 0.4 0.5 0.4 0.8 0.4 0.4 0.5 0.4 0.6 0.4 0.3 0.3 0.3 0.4 0.2 0.5 0.4 2.3 2.4 2.4 2.1 1.5 1.5 3.4 3.6 2.9 2.9 1.9 2.5 1.8 3.8 1.9 1.8 2.7 1.8 3.2 1.9 1.4 1.6 1.6 2.2 1.0 2.6 1.8 9.7 10.0 10.2 12.2 9.8 12.9 13.8 16.8 10.7 15.3 9.2 11.5 7.0 17.8 8.7 7.8 11.0 8.7 15.7 8.8 7.2 7.4 7.8 10.0 4.8 9.7 7.0 22.7–24.6 22.8–24.8 22.9–24.9 16.5–18.2 14.2–15.4 11.2–12.5 23.3–26.1 19.7–22.6 25.9–28.3 17.8–20.1 20.2–21.8 20.6–22.7 25.5–27.0 20.0–23.1 20.6–22.1 21.6–23.0 23.4–25.7 20.4–21.9 19.3–22.0 20.4–21.9 18.8–20.0 20.3–21.6 20.4–21.7 21.3–23.1 20.9–21.7 25.7–27.8 25.4–26.9 Note: Arithmetic average (x), average standard deviation (sx), sample standard deviation (s), coefficient of variability (V%) and 95% confidence interval (CI). tectate, thicker than nexine, and may be slightly thicker at one of the poles (Figure 2A, Table IV). In general, Dioscorea pollen grain size and exine ornamentation patterns were found to be highly variable, allowing the distinction of two species and groups of species. Principal component analysis (PCA) The pollen variability among the species and specimens amounted to 89.05% on the two first axis in the PCA (Figure 6). The first axis summed 77.03% of the variation, according to the highly correlated values (> 0.23), especially the larger equatorial axis in polar view (VPEM), shorter equatorial axis in polar view (VPEm), equatorial axis in equatorial view (VEEM) and the polar axis in equatorial view (VEEm) (Table V). The second component explained 12.02% of the variability of the data, in which the length of sulculi (SUCC), the thickness of the sexine layer (SEX) and thickness of the nexine layer (NEX) were the most important variables (> 0.12) for the ordination of specimens. On the left side of the graph (Figure 6) were ordered the specimens related to the larger pollen grains. High values of the exine thickness were important for the specimen ordination on the top left side of the graph. In contrast, the specimens ordered on the right side of the graph showed the smaller dimensions for these variables, with three specimens of Dioscorea amaranthoides being the smallest pollen grains. Cluster analysis (UPGMA and Euclidean distance) The CA provided a similarity dendrogram for the studied specimens by calculation of the Euclidean distance (Figure 7). In this dendrogram, it was possible to identify the presence of two groups of specimens with 0% similarity, the first comprising the specimens of Dioscorea amaranthoides, and the second accommodating the rest of the species, separating the smallest pollen grains (specimens from the first group) from the largest pollen grains (specimens from the second group). Dioscorea olfersiana (SP 377050) and D. tauriglossum (SP 472163) presented a 100% similarity, and Pollen morphology of Dioscorea 5 Table III. Measures (in micrometres) in equatorial view of Dioscorea L. pollen grains through light microscopy (n = 25). Equatorial axis in equatorial view (VEEM) Species Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea altissima (SP 304317) altissima (SP 446562) altissima (SP 487481) amaranthoides (SP 49534) amaranthoides (SP 168574) amaranthoides (SP 446580) dodecaneura (SP 298281) dodecaneura (SP 246876) dodecaneura (SP 276042) laxiflora (SP 202296) laxiflora (SP SP 248709) laxiflora (SP 287969) monadelpha (SP 426529) monadelpha (SP 472164) monadelpha (SP 487480) multiflora (SP 289468) multiflora (SP 444406) multiflora (SP 475315) olfersiana (SP 318370) olfersiana (SP 377050) olfersiana (SP 474884) tauriglossum (SP 248719) tauriglossum (SP 472163) tauriglossum (SP 472166) trilinguis (SP 53704) trilinguis (SP 212491) trilinguis (SP 426533) Polar axis in equatorial view (VEEm) x sx s V% CI x sx s V% CI 34.0 28.2 29.5 22.0 20.7 22.0 34.9 36.1 36.0 28.4 27.7 28.8 30.8 27.2 29.4 33.6 30.4 30.2 28.5 25.4 25.1 27.1 25.9 29.7 29.9 30.7 30.3 1.1 0.6 0.4 0.3 0.2 0.4 0.4 0.7 0.5 0.6 0.3 0.3 0.4 0.4 0.4 0.5 0.5 0.4 0.4 0.3 0.3 0.4 0.4 0.4 0.2 0.3 0.2 5.6 3.0 2.1 1.4 1.2 1.9 1.9 3.5 2.7 2.8 1.4 1.5 1.9 2.1 1.8 2.5 2.3 2.2 2.0 1.6 1.5 1.8 1.8 2.0 1.2 1,5 1.1 16.3 10.4 7.2 6.1 6.0 8.5 5.4 9.7 7.5 9.8 5.2 5.3 6.1 7.8 6.2 7.6 7.6 7.4 6.9 6.3 5.8 6.8 6.8 6.9 4.0 4.9 3.6 31.7–36.3 27.0–29.5 28.6–30.4 21.5–22.6 20.2–21.2 21.2–22.8 34.1–35.6 34.6–37.5 34.8–37.1 27.3–29.5 27.1–28.3 28.2–29.4 30.1–31.6 26.3–28.1 28.6–30.1 32.6–34.7 29.4–31.3 29.2–31.1 27.7–29.3 24.8–26.1 24.5–25.7 26.4–27.9 25.2–26.6 28.9–30.6 29.4–30.4 30.1–31.3 29.8–30.7 22.8 22.2 23.7 15.5 14.1 15.5 23.1 20.6 26.1 16.9 20.2 20.0 25.6 21.3 19.9 21.0 27.7 20.4 20.4 19.0 19.4 20.2 19.3 20.4 21.5 22.3 22.3 0.5 0.5 0.4 0.4 0.4 0.4 0.6 0.5 0.6 0.5 0.3 0.5 0.4 0.5 0.4 0.4 0.5 0.5 0.3 0.4 0.4 0.3 0.3 0.4 0.2 0.3 0.2 2.6 2.6 2.1 1.8 1.8 1.8 3.2 2.5 3.2 2.5 1.6 2.5 2.0 2.6 2.2 2.0 2.3 2.5 1.4 2.0 1.8 1.4 1.7 2.0 1.0 1.7 1.1 11.5 11.8 8.8 11.7 13.0 11.8 13.9 12.1 12.1 15.0 7.8 12.4 7.6 12.4 11.1 9.4 9.9 12.2 6.8 10.7 9.3 6.8 9.0 9.6 4.5 7.5 5.1 21.7–23.9 21.1–23.3 22.9–24.6 14.8–16.3 13.3–14.9 14.7–16.2 21.8–24.4 19.6–21.7 24.8–27.4 15.8–17.9 19.6–20.9 19.0–21.0 24.8–26.4 20.2–22.3 19.0–20.8 20.2–21.8 21.8–23.7 19.4–21.4 19.9–21.0 18.1–19.8 18.7–20.1 19.7–20.8 18.6–20.0 19.6–21.2 21.1–21.9 21.6–23.0 21.8–22.7 Note: Arithmetic average (x), average standard deviation (sx), sample standard deviation (s), coefficient of variability (V%) and 95% confidence interval (CI). D. laxiflora (SP 248709) and D. tauriglossum (SP 248719) amounted to almost 100% similarity. In general, the specimens of each species showed values of pollen similarity below 50%, indicating that there is significant variation in the pollen characteristics of these species. The exception was Dioscorea amaranthoides and D. dodecaneura specimens, which grouped together both in PCA and CA. Pollen types Four pollen types have been recognised based on pollen exine ornamentation. Five pollen subtypes were additionally recognised, based on the grain size. Pollen type 1. — Pollen with striate exine ornamentation Subtype 1a. — With narrow striae running in two directions, and tightly connected striations, small sized (longest axis reaching 16.7 to 22.0 µm) (Dioscorea amaranthoides). Subtype 1b. — With narrow striae running in two directions, and tightly connected striations, predominantly medium sized, with rare small sized pollen grains (longest axis reaching 24.5 to 28.9 µm) (Dioscorea laxiflora, D. olfersiana). Comments. — Only the VPEM average values of Dioscorea laxiflora SP 202296 and SP 287969 exceeded these values. Subtype 1c. — With broader striae running only in one direction, and not so tightly connected striations, medium sized (longest axis reaching 29.2 to 36.6 µm) (Dioscorea trilinguis). Comments. — Only the VPEM average values of Dioscorea trilinguis SP 53704 were lower. Pollen type 2. — Pollen with microreticulate exine ornamentation Subtype 2a. — Predominantly medium sized, with rare small sized pollen grains (longest axis reaching 24.5 to 28.9 µm) (Dioscorea tauriglossum). 6 C. F. P. da Luz et al. Figure 1. Light microscopy (LM) images of Dioscorea pollen grains. A–E. Dioscorea altissima Lam. A. Polar view, optical section (SP 446562). B. Polar view, optical section (SP 487481). C. Equatorial view, surface and sulcus details (SP 487481). D. LO1 (high focus) (SP 487481). E. LO2 (low focus) (SP 487481). F–J. Dioscorea amaranthoides C. Presl. (SP 49534). F. Polar view, optical section. G. Equatorial view, optical section. H. Equatorial view, surface and aperture details. I. LO1 (high focus). J. LO2 (low focus). K, L. Dioscorea dodecaneura Vell. SP 276042. K. Polar view, optical section. L. Polar view, surface. Scale bars – 10 µm (A–C, F–H, K, L), 2 µm (D, E, I, J). Comments. — Only the VEEM average values of Dioscorea tauriglossum SP 472166 exceeded these values. Pollen type 3. — Pollen with psilate-perforate exine ornamentation, medium sized (longest axis reaching 29.2 to 36.6 µm) (Dioscorea altissima). Subtype 2b. — Medium sized (longest axis reaching 29.2 to 36.6 µm) (Dioscorea dodecaneura, D. multiflora). Pollen type 4. — Pollen with rugulate-perforate exine ornamentation, medium sized (longest axis reaching 29.2 to 36.6 µm) (Dioscorea monadelpha). Pollen morphology of Dioscorea 7 Figure 2. Light microscopy (LM) images of Dioscorea pollen grains. A–C. Dioscorea dodecaneura Vell. SP 276042. A. Equatorial view, optical section. B. LO1 (high focus). C. LO2 (low focus). D–H. Dioscorea laxiflora Mart. ex Griseb. SP 287969. D. Polar view, surface. E. Equatorial view, optical section. F. Equatorial view, surface and aperture details. G. LO1 (high focus). H. LO2 (low focus). I–L. Dioscorea monadelpha (Kunth) Griseb. I. Polar view, optical section (SP 472164). J. Polar view, surface (SP 472164). K. Equatorial view, optical section (SP 472 164). L. LO1 (high focus) (SP 426529). Scale bars – 10 µm (A, D–F, I–K), 2 µm (B, C, G, H, L). Comments. — Only the VPEM and VEEM average values of Dioscorea monadelpha SP 472164 were lower. Discussion In the present work, only one aperture pattern was documented, therefore it was not useful for recognis- ing species groups; however, pollen size and exine ornamentation were important characteristics for the typification of pollen groups. Four pollen types (and five subtypes) were recognised based on these palynological characters. Caddick et al. (1998) had already concluded that both microsporogenesis and pollen morphology contribute with taxonomically 8 C. F. P. da Luz et al. Figure 3. Light microscopy (LM) images of Dioscorea pollen grains. A. Dioscorea monadelpha (Kunth) Griseb. LO2 (low focus) (SP 426529). B–H. Dioscorea multiflora Mart. ex Griseb. SP 444406. B. Polar view, optical section. C. Polar view, surface. D. Equatorial view, surface and aperture details. E. Frontal equatorial view, optical section. F. Frontal equatorial view, surface. G. LO1 (high focus). H. LO2 (low focus). I–L. Dioscorea olfersiana Klotzsch ex Griseb. I. Polar view, optical section (SP 474884). J. Equatorial view, optical section (SP 474884). K. Equatorial view, surface (SP 474884). L. LO1 (high focus) (SP 318370). Scale bars – 10 µm (B–F, I–K), 2 µm (A, G, H, L). informative characters at genus level, especially in Dioscoreaceae, in relation to their closest allies, but are probably less informative at higher taxonomic levels. Species delimitation was not possible based on pollen metric parameters, since a wide range of variation was documented among specimens of the same species, generating a degree of overlap. Standardisation of terminology An issue encountered in the present study was the disparity of terminology used in the consulted literature to describe different aspects of the pollen morphology, which made it difficult to establish an accurate comparison of previous results with our data. We provide in Table VI a comparison of the Pollen morphology of Dioscorea 9 Figure 4. Light microscopy (LM) images of Dioscorea pollen grains. A. Dioscorea olfersiana Klotzsch ex Griseb. SP 318370. LO2 (low focus). B–F. Dioscorea tauriglossum R. Knuth. SP 472163. B. Polar view, optical section. C. Polar view, surface. D. Equatorial view, optical section. E. LO1 (high focus). F. LO2 (low focus). G–L. Dioscorea trilinguis Griseb. G. Polar view, optical section (SP 426533). H. Polar view, surface (SP 426533). I. Equatorial view, optical section (SP 212491). J. Equatorial view, surface (SP 212491). K. LO1 (high focus) (SP 212491). L. LO2 (low focus) (SP 212491). Scale bars – 10 µm (B–D, G–J), 2 µm (A, E, F, K, L). terminology used by previous authors with the one we have applied, which follows Punt et al. (2007) for standardisation. For instance, we have often found the use of ‘colpus’ or ‘sulcus’, which should have been categorised as sulculi, since they are situated between the proximal and distal poles (‘an elongated latitudinal ectoaperture not situated at a pole’); different variation in pollen size; diverse terminology 10 C. F. P. da Luz et al. Figure 5. Scanning electron mcroscopy (SEM) images of Dioscorea pollen grains. A–B. Dioscorea altissima Lam. SP 304317. A. Equatorial view, two pollen grains. B. Psilate-perforate sexine ornamentation. C, D. Dioscorea amaranthoides Presl. SP 168574. C. Equatorial view, two pollen grains. D. Distal polar view. E, F. Dioscorea dodecaneura Vell. SP 276042. E. Distal polar view. F. Microreticulate sexine ornamentation. G. Dioscorea laxiflora Mart. ex Griseb. SP 287969. Equatorial view. H, I. Dioscorea monadelpha (Kunth) Griseb. SP 426529. H. Equatorial view. I. Rugulate-perforate sexine ornamentation. J, K. Dioscorea multiflora Mart ex Griseb. SP 444406. J. Distal polar view. K. Microreticulate sexine ornamentation. L. Dioscorea olfersiana Klotzsche ex Griseb. SP 318370. Proximal polar view. M. Dioscorea tauriglossum R. Knuth. SP 472163. Equatorial view. N, O. Dioscorea trilinguis Griseb. SP 212491. N. Distal polar view. O. Striate sexine ornamentation. Scale bars – 10 µm (A, E, J, N), 5 µm (C, D, G, H, L, M, O), 2 µm (B, F, I, K). Pollen morphology of Dioscorea 11 Table IV. Arithmetic average of measures (in micrometres) of the sulcus length and thickness of the exine layers of the Dioscorea L. pollen grains through light microscopy (n = 10). Species SUCC SEX NEX Exine ornamentation 28.00 24.41 21.48 18.78 0.93 0.67 0.79 0.75 0.83 0.57 0.69 0.62 Dioscorea amaranthoides (SP 168574) 14.19 0.72 0.62 Dioscorea amaranthoides (SP 446580) 16.34 0.70 0.48 dodecaneura (SP 298281) dodecaneura (SP 246876) dodecaneura (SP 276042) laxiflora (SP 202296) 30.09 31.56 27.14 24.23 0.89 0.92 0.90 0.80 0.79 0.82 0.80 0.69 Dioscorea laxiflora (SP SP 248709) 19.93 0.84 0.75 Dioscorea laxiflora (SP 287969) 23.47 0.89 0.79 22.02 18.40 20.23 26.66 21.61 21.52 20.35 1.03 0.84 0.90 0.88 0.93 0.86 0.96 0.92 0.74 0.81 0.80 0.83 0.79 0.86 Dioscorea olfersiana (SP 377050) 21.23 0.83 0.74 Dioscorea olfersiana (SP 474884) 20.73 0.76 0.67 21.11 21.61 21.27 21.13 0.87 0.87 0.94 0.89 0.77 0.77 0.84 0.79 Dioscorea trilinguis (SP 212491) 22.72 0.99 0.89 Dioscorea trilinguis (SP 426533) 22.20 0.90 0.80 Psilate-perforate Psilate-perforate Psilate-perforate Striate, with narrow striae running in two directions, and tightly connected striations Striate, with narrow striae running in two directions, and tightly connected striations Striate, with narrow striae running in two directions, and tightly connected striations Microreticulate Microreticulate Microreticulate Striate, with narrow striae running in two directions, and tightly connected striations Striate, with narrow striae running in two directions, and tightly connected striations Striate, with narrow striae running in two directions, and tightly connected striations Rugulate-perforate Rugulate-perforate Rugulate-perforate Microreticulate Microreticulate Microreticulate Striate, with narrow striae running in two directions, and tightly connected striations Striate, with narrow striae running in two directions, and tightly connected striations Striate, with narrow striae running in two directions, and tightly connected striations Microreticulate Microreticulate Microreticulate Striate, with broader striae running rather only in one direction, and not so tightly connected striations Striate, with broader striae running rather only in one direction, and not so tightly connected striations Striate, with broader striae running rather only in one direction, and not so tightly connected striations Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea Dioscorea altissima (SP 304317) altissima (SP 446562) altissima (SP 487481) amaranthoides (SP 49534) monadelpha (SP 426529) monadelpha (SP 472164) monadelpha (SP 487480) multiflora (SP 289468) multiflora (SP 444406) multiflora (SP 475315) olfersiana (SP 318370) tauriglossum (SP 248719) tauriglossum (SP 472163) tauriglossum (SP 472166) trilinguis (SP 53704) Note: SUCC, sulcus length; SEX, thickness of the sexine layer; NEX, thickness of the nexine layer. applied to the characterisation of the type of ornamentation of the exine. Barroso et al. (1974), having on his team palynologist Dr Ortrud Monika Barth, conducted a survey of the Flora of Guanabara (Rio de Janeiro, Brazil), with LM analysis of the pollen grains of 24 species of Dioscorea; they have been described as dicolpate and were grouped into four pollen types, according to the exine ornamentation, among which five of the species studied here. The descriptions provided by the authors contemplate striate exine in Dioscorea laxiflora, D. olfersiana and D. trilinguis, and reticulate in D. dodecaneura (here described as microreticulate), corroborating our results. However, the reticulate-ornamented exine of D. altissima (syn. D. chondrocarpa) differs from our psilate-perforate description. Corrêa (1996) studied, under LM, five species of Dioscorea present in the ‘Parque Estadual das Fontes do Ipiranga’ (São Paulo, Brazil), including four of the analysed in the present work. All of them were characterised as dicolpate. The ornamentation of D. dodecaneura and D. multiflora has been characterised as microreticulate, corroborating the present work. Meanwhile, Dioscorea olfersiana was characterised as reticulate-striate, whereas here the ornamentation was only striate, and D. monadelpha was characterised by the author as microreticulate, and in the present study, with use of scanning electron microscopy (SEM), it was verified that it is rugulate-perforate. 12 C. F. P. da Luz et al. Figure 6. Principal component analysis (PCA) performed with the pollen metrical variables from Dioscorea. 1. D. altissima; 2. D. amaranthoides; 3. D. dodecaneura; 4. D. laxiflora; 5. D. monadelpha; 6. D. multiflora; 7. D. olfersiana; 8. D. tauriglossum; 9. D. trilinguis. Table V. Pearson and Kendall correlation coefficients for pollen grains metric variables of the first and the second axis of principal component analysis (PCA) ordination in Dioscorea L. Principal components Variables VPEM VPEm VEEM VEEm SEX NEX SUCC Variables (meaning) Axis 1 Axis 2 Larger equatorial axis in polar view Shorter equatorial axis in polar view Equatorial axis in equatorial view Polar axis in equatorial view Sexine thickness Nexine thickness Length of sulcus −0.4382 −0.4537 −0.3914 −0.3650 −0.2378 −0.3331 −0.3851 −0.1685 0.2353 −0.2266 0.1293 0.3983 0.5043 −0.6597 Xifreda (2000) palynologically analysed 45 species, and one variety, of Dioscorea, across 25 sections from South America, including seven species here studied, with exception of D. laxiflora and D. tauriglossum. The author described the pollen of D. amaranthoides and D. olfersiana as striate, corroborating our results. The pollen of D. altissima (syn. D. chondrocarpa) was described as perforate-scrobiculate, an outdated term for perforate ornamentation, therefore, also in agreement with our results. However, some descriptions by this author did not correspond to the ones here observed. Dioscorea monadelpha, with pollen described as rugulate-reticulate, did not correspond to the rugulate-perforate in the present work. According to Punt et al. (2007), a reticulum is: ‘A network-like pattern consisting of lumina or other spaces wider than 1 μm bordered by elements narrower than the lumina’, which is not what can be observed in SEM images presented by the author, to describe the pollen of this species. The images presented by Xifreda (2000) for the rugulate-reticulate pollen type is, in fact, better described as rugulate-perforate ornamentation, with differences in the width and foldings of the rugulae, and in the presence or absence of perforations between them or above them. In D. trilinguis, the perforatescrobiculate description of the material analysed by the author differed greatly from the striate perforation of the specimens in the present study, as well as the description of D. dodecaneura and D. multiflora as perforate-scrobiculate, here found to be microreticulate. This variation in the exine ornamentation is unusual between specimens of the same species (SalgadoLabouriau et al. 1965), and the correct specimen identification should be investigated by taxonomists. Schols et al. (2001) studied the pollen morphology of 35 species of Dioscorea using only SEM and Pollen morphology of Dioscorea 13 Figure 7. Cluster analysis (UPGMA, Euclidean distance) performed with the pollen metrical variables of Dioscorea species. Abbreviations: see caption of Figure 6. transmission electron microscopy (TEM), including only D. dodecaneura and D. multiflora, out of all the species here studied. They described D. dodecaneura as monosulcate or, more frequently, disulcate, of small to medium size and perforate ornamentation; and D. multiflora as disulcate, of small to medium size, and perforate ornamentation. The disulcate pollen was predominant in the species of Dioscorea examined by the authors; however, monosulcate was predominant in sections Stenophora and Monadelpha. In the two species that were also analysed by us, the pollen grains were only found to be disulculate, of medium size, with microreticulate ornamentation, which did not coincide with the author descriptions. Schols et al. (2003) studied more extensively the pollen grains of the family, including 61 species of Dioscorea, covering seven Brazilian species (D. anomala Griseb., D. debilis Uline ex Knuth, D. dodecaneura Vell., D. glandulosa Klotzsch ex Kunth, D. olfersiana Klotzsch ex Griseb., D. piperifolia Klotzsch ex Kunth, D. sinuata Lam.). In this work, the authors corrected the classification of the pollen aperture type of Dioscorea given by Schols et al. (2001), from sulcate to sulculate, after confirmation under TEM of the pollen of D. communis. To these authors, the pollen grains of D. dodecaneura are of small to medium size, disulculate and perforate, partly agreeing with our results, since only medium-sized and microreticulate pollen grains were observed. To these authors, D. olfersiana presents monosulcate or disulculate, and striate, pollen grains; here, only disulculate pollen grains occurred, which striate ornamentation corresponding to the one described by them. Alzer (2017) studied 24 species of Dioscorea L., 19 of which of Dioscorea subg. Helmia, and five of D. subg. Dioscorea, delimitating seven pollen types based on exine ornamentation. The author studied four species, which were also here analysed: D. altissima (syn. D. chondrocarpa), D. monadelpha, D. multiflora and D. olfersiana, all of which were described as disulcate. It is noteworthy that the apertures characterised by the author as sulcate do not correspond to the ones analysed under TEM by Schols et al. (2005a) and described as sulculate. The ornamentation in D. altissima was described by the author as perforate, and in D. olfersiana as striate, supporting our results. For D. monadelpha, the author described the exine as reticulate-cristate (which reticula muri presented prominent sculptural elements) and triangular amb, characteristics which were not observed in any specimen here analysed. The exine of D. multiflora was described by the author as reticulate, however in our material the lumina in the reticula network are smaller than 1 µm, hence being described by us as microreticulate. According to Schols et al. (2005b) during the ontogeny of Dioscorea pollen, the exine development starts at the proximal pole, and during the tetrad stage, the exine remains thicker at this pole, which sculptural elements can be thicker and more developed there. This fact was visually verified in D. dodecaneura, here analysed. However, the other authors cited in Table VI did not highlight this characteristic for this species. Still according to Schols et al. (2005b), the development sites for two sulculi are equatorial, positioned between the proximal and distal poles, forming an angle of 180°, sometimes slightly shifted from the centre. This fact was observed here in several taxa, which sulculi were not perfectly positioned in the centre of the pollen grain. To these authors, the pollen grains have a curved proximal pole in outline, and a plane distal pole. In our material, it was not always possible to verify this distinction, the outline being characterised as biconvex in most species, but we have observed that, in many cases, the area of one pole was smaller than the other. Different exine ornamentation patterns were also described for Dioscorea species which were not here analysed, as, for instance, psilate, reticulate or scabrate, by Sharma (1967); granulate, by Huang 14 Barroso et al. (1974) Total of species studied Shape of pollen grains Size of pollen grains Aperture type Number of apertures Sexine sculpturing Corrêa (1996) Xifreda (2000) 24 (24 specimens) 5 (11 specimens) Oval Oval — Small to medium sized (15.0– 25.0 µm) Colpus Medium sized (25.6–31.3 µm) – Colpus 45 (48 specimens) Colpus 2 2 2 Reticulateornamentated (D. altissima), reticulate (D. dodecaneura), striate (D. laxiflora, D. olfersiana, D. trilinguis) Microreticulate (D. dodecaneura, D. monadelpha, D. multiflora), reticulate-striate (D. olfersiana) Perforate-scrobiculate (corresponding to perforated ornamentation) (D. altissima, D. dodecaneura, D. multiflora, D. trilinguis), striate (D. amaranthoides, D. olfersiana a), rugulatereticulate (D. monadelpha) Schols et al. (2001) Schols et al. (2003) 35 (35 specimens) — 61 (61 specimens) — Small to medium sized (15– 51 µm) Sulcus Small to medium sized (20– 58 µm) Sulculus or sulcus 2 (1) 2 (1) Perforate, 2 (1) sulcus (D. dodecaneura), perforate, 2 sulcus (D. multiflora) Perforate, 2 sulculus (D. dodecaneura), striate, 2 sulculus, less frequent 1 sulcus (D. olfersiana) Alzer (2017) This study 24 (50 specimens) 9 (27 specimens) Elliptic or elliptictriangular amb Small to medium sized (14.8– 35.5 µm) Sulcus Spheroidal to ellipsoidal Small to medium sized (16.7–36.6 µm) Sulculus 2 (1) 2 Perforate, 2 sulcus (D. altissima), reticulate-cristate with elliptictriangular amb, 2 sulcus (D. monadelpha), reticulate, 2 sulcus (D. multiflora), striate, 2 sulcus (D. olfersiana) Psilate-perforate (D. altissima), striate (D. amaranthoides, D. laxiflora, D. olfersiana, D. trilinguis), microreticulate (D. dodecaneura, D. multiflora, D. tauriglossum), rugulate with perforations (D. monadelpha) C. F. P. da Luz et al. Table VI. Summary of aperture features and descriptions of pollen grains sexine ornamentation of Dioscorea L. species obtained in this research (last column), compared to the results of same species according to other authors. Pollen morphology of Dioscorea (1970); brain-like reticulate by Su (1987); scabrate by Roubik and Moreno (1991); supra-reticulate (bireticulate) by Heusser (1971); gemmate by Caddick et al. (1998); brain-like perforate (strongly ondulate with two types of perforation) by Wilkin et al. (2002); medium-size striate, which striae are arranged in concentrical polygons, or not as markedly concentrical, by Wilkin et al. (2009); with internal granules in the lumina, and the presence of smaller lumina around the larger lumina, by Couto et al. (2014). Aperture number The pollen grains of Dioscorea are released individually as monads and are usually monosulcate or, most frequently, disulcate, and rarely triaperturate (Caddick et al. 1998; Schols et al. 2001). A specimen of D. densiflora Hemsl. (Bourgeau 7687), from Mexico, was described by Erdtman (1952) as disulculate or, occasionally, trisulculate. The pollen of D. deltoidea Wall. has been described as tricolpate, occasionally dicolpate, and that of D. melanophyma Prain & Burkill as dicolpate, occasionally tricolpate (Gupta & Sharma 1986). The pollen of D. bulbifera L. has been decribed as 2–3 anasulcate (sulcus on the distal pole) (Su 1987). Although the monosulcate pollen type has been reported in other studies, it was not found in ours, where disulculate pollen is shared by all species analysed, which have been resolved in three different clades in the phylogeny of Couto et al. (2018) – New World I, New World II and New World III – and have been classified in eight different sections by Knuth (1924). Therefore, even with a relatively small taxonomic sampling of nine species, our results suggest that the number of apertures is likely not a character of systematic value among the Neotropical species of Dioscorea. 15 Pollen size reported in our analyses differed from all the bibliographic references consulted, either being larger or smaller than the dimensions observed here. Probably, most of the differences are due to the methodology applied, since some authors used the pollen preparation by lactic acetolysis by Raynal and Raynal (1971), while others have used Erdtman’s (1960) classic acetolysis. In our study, we have reported variation in pollen size among specimens of the same species, which suggests that pollen size is not a suitable character to differentiate species. However, the pollen grains of Dioscorea amaranthoides were the smallest and could be separated from the remaining species. The largest pollen size was found in D. dodecaneura. Dioscorea monadelpha is the only monoecious species of the taxa here analysed, with the staminate and pistillate inflorescences occurring on the same specimen (Xifreda & Kirizawa 2003). The dioecy of most taxa in our study has made it difficult to collect specimens with staminate flowers, in areas traversed by the road ‘Rodoanel Mário Covas’, which has led us to include additional material from other localities within the States of São Paulo and Paraná, to be able to fully characterise each species. The wider geographical range of the collection might account for the differences in pollen grain size, observed between specimens of the same species, since in distinct environmental conditions, different localities, and among populations, variation in pollen size may occur (Jones & Newelll 1942; Cain & Cain 1948; Martin 1959). These results show the importance of verifying to what extent measurements really represent the species (Salgado-Labouriau et al. 1965) and, hence, the importance of producing measurements across different specimens. Exine ornamentation Pollen size Su (1987) has suggested that the size of pollen grains in Chinese species of Dioscorea correlates with the type of tuber. Schols et al. (2001, 2003) and Caddick et al. (2002) have confirmed this correlation in pollen grain size with the type of tuber found in some species of Dioscorea. The cited authors have stated that this correlation in pollen size with the tuber type could well be linked to genetic characteristics. Thus, Schols et al. (2005a) have indicated that, in Dioscorea, the pollen size decreases in the most derived clades, and could be related to the evolution of annually replaced tubers. Among the species here analysed, the tuber of D. amaranthoides was not the smallest, but its pollen was, challenging the hypothesis proposed by Su (1987). The description of the ornamentation of the exine varied significantly, also among the cited authors. However, only a few of the consulted studies have utilised SEM, therefore the accuracy of some of the descriptions is difficult to be relied on. The most subtle variations in ornamentation, such as the presence of perforations, can only be observed in SEM, and are virtually impossible to be distinguished in LM. The thickness of sexine and nexine, however, can be observed under LM, hence our data could be compared with previously published studies. In all works, including ours, the sexine was found to be thicker than the nexine. Four different types of exine ornamentation were here recorded, which were the main charac- 16 C. F. P. da Luz et al. ter for the recognition of pollen types in this study: striate (in one and two directions), microreticulate, psilate-perforate (only present in Dioscorea altissima) and rugulate-perforate (only present in D. monadelpha). Environmental conditions Out of the nine species of Dioscorea here analysed, eight occupy the low-woodland of the Dense Ombrophilous Forest of the Serra do Mar, except for D. amaranthoides, collected in anthropized Cerrado. It is noteworthy that the pollen of this species was the smallest of all, which could correlate to the drier and hotter Cerrado environmental conditions, since, among other factors, pollen size depends on the level of hydration of the anther (Halbritter et al. 2018). Ejsmond et al. (2015), in studying 232 species of 11 taxonomic groups, corroborated that the interspecific variation in pollen size is driven by environmental conditions during the flowering season, especially temperature. Implications for the systematics of Neotropical species of Dioscorea The pollen types that we have recognised here based on exine ornamentation corroborate in part the species relationships suggested by Couto et al. (2018). Pollen type I, striate (Dioscorea amaranthoides/D. laxiflora/D. olfersiana/D. trilinguis). — Clade New World I as defined by Couto et al. (2018), contains four of the species here analysed – Dioscorea amaranthoides, D. laxiflora, D. olfersiana and D. multiflora. Three of these presented this pollen type, with striate exine, running in two directions (D. amaranthoides, D. laxiflora, D. olfersiana), while D. multiflora was placed in a different pollen type (Pollen type 2), possessing microreticulate exine. New World I is a large clade that is subdivided into other smaller clades. It is noteworthy that, eventhough all four species belong in New World I, D. multiflora belongs in a separate clade from the other three species, which could account for the evident differences in its exine. Both D. laxiflora and D. olfersiana present sessile staminate flowers and six stamens fixed to the half of the perigonium tube, and, therefore, have been placed in the same section (Section Cryptantha Uline.) by Knuth (1924). The molecular phylogenetic results and our palynological analysis further confirm their close relationship. In the pollen analysis, they presented the same pollen subtype, encompassing medium-sized pollen grains (24.5– 28.9 µm on the longest axis) with striate exine. The UPGMA results have also reinforced the proximity between the two species, where specimens from both species presented over 85% similarity. Dioscorea amaranthoides was the most distinctive of all four species, with the UPGMA and PCA results suggesting 0% similarity with the remaining of the group, and the main characters contributing to that separation being the much smaller size of the pollen grains and also of the smaller wall thickness. The species Dioscorea trilinguis, which has not yet been sampled in a molecular phylogenetic analysis, also presented striate exine, although running in one direction (and not two). It is possible that it also belongs in the clade New World I, close to D. amaranthoides, D. laxiflora and D. olfersiana. In the UPGMA, two specimens of D. trilinguis presented low similarity (c. 35%) with D. olfersiana and D. laxiflora. Only one specimen of D. trilinguis presented higher similarity with these species (62%). The characters most contributing to this separation where the exine ornamentation and the metric variables, which additionally presented infra-specific variation. Pollen type 2, microreticulate (Dioscorea dodecaneura/ D. multiflora/D. tauriglossum). — was resolved in New World III (Couto et al. 2018), while D. multiflora was resolved in New World I, in a clade that is separate from the clade of D. amaranthoides, D. laxiflora and D. olfersiana. The clades are well supported and, unless there are uncertainties about the correct identification of the samples, the results of Couto et al. (2018) point out to a distance relationship between the two species. Morphologically, they are also very distinct. Dioscorea dodecaneura possesses pedicellate flowers and six stamens fixed at the base of the perigonium tube, while D. multiflora possesses sessile flowers, or almost, with six stamens fixed at the base of the perigonium tube. The pollen grains of the specimens of Dioscorea dodecaneura have presented almost 70% similarity among them, while those of D. multiflora have presented great intraspecific variation, making it difficult to characterise this species in terms of its pollen dimensions. It is possible that the occurrence of the varieties synonymised by Xifreda (1982), included in this species, is what accounts for the difficulty in the palynological characterisation of the species. In addition, in the State of São Paulo, D. multiflora occurs in Dense Ombrophilous Forest, as well as in the Semi-deciduous Forest, a less humid environment. The possession of this pollen type may not signify a close relationship with Dioscorea dodecaneura and/ or D. multiflora, since these two species are already known to be distantly related. Morphologically, D. Pollen morphology of Dioscorea multiflora shares with D. dodecaneura the possession of pedicellate flowers, differing from it for having leaf blade elliptic, lanceolate or elliptc-lanceolate, with five nerves and 1–4 staminate flowers, in a scorpioid cyme. However, in D. dodecaneura, the leaves are cordate, with 9–10 veins and staminate flowers, solitary and in D. multiflora, the flowers are sessile, or almost, grouped into each floral node, and the leaf blade varies from cordate to ovate, to ovateelliptic, with 5–7 veins. Our results for Dioscorea tauriglossum are original. Not only has it never been palynologically analysed in previous studies, but also it has not been sampled in any of the molecular phylogenetic studies available. Given that the two other species in this pollen type are not phylogenetically closely related (New World I and New World III), it is difficult to assume in which clade Dioscorea tauriglossum could be resolved. Looking at the distance relationship between D. dodecaneura and D. multiflora, it seems that the possession of this exine ornamentation would be a result of convergent evolution, and might not be of systematic value. Pollen type 3, exine psilate-perforate (Dioscorea altissima, syn. D. chondrocarpa). — Only Dioscorea altissima (D. chondrocarpa) was found to possess this type of exine ornamentation. This species has not yet been sampled in a molecular phylogenetic study, so it is not possible to determine to which species it is more closely related to. Since no other species here analysed, from any of the clades resolved in the phylogeny of Couto et al. (2018) – New World I, New World II, New World III – possess this type of exine, it is possible that it is not closely related to any of them, or that this is a derived trait, within one of such clades. Pollen type 4, exine rugulate-perforate (Dioscorea monadelpha). — All species here analysed are dioecious and present free styles, in general with introrse anthers, except for Dioscorea monadelpha, which possesses a staminal column, longitudinal dehiscence and winged seeds, which depend on wind for dispersal. Dioscorea monadelpha was the only species to present rugulate-reticulate ornamentation. Xifreda (2000) has documented this type of exine ornamentation in other South American species of Dioscorea, all belonging in subgenus Helmia, like D. monadelpha, but distributed among three different sections: reported in all species of Section Centrostemon and some species of Section Dematostemon and Section Monadelpha. Of all the species in Xifreda (2000) found to possess this type 17 of exine ornamentation, three were included in the phylogeny of Couto et al. (2018) – D. grandiflora in Section Cetrontrostemon, and D. coronata and D. monadelpha, in Section Monadelpha – which were all resolved in the clade New World II. Since this trait has not been reported in species outside of clade New World II, it is possible that exine ornamentation is of systematic value for characterising this clade, or groups within it. Conclusions The exine ornamentation features and the size of the pollen grains were variable, allowing its use in the recognition of different pollen types. In general, however, pollen morphological characteristics suggested groupings of unrelated species, in the sense that exine ornamentation and metric parameters did not reflect the separation of the species into the already recognised sections. However, under the light of recent evolutionary and phylogenetic studies of Dioscorea, our results help to support some phylogenetic relationships between species and redflag groups where characters may or may not be of systematic value (for instance, in the pollen type that contains species from different clades). We hope that the present palynological data will contribute to the improvement of the pollen morphological knowledge of Dioscorea, particularly the Neotropical species, and can be used for future phylogenetic approaches to solve taxonomic problems, contributing to the progress in systematic studies of yams. Acknowledgements The authors are grateful to the curator of the Herbarium SP that helpfully allowed access to herbarium specimens for sampling. Thanks are also due to Luciana Benjamin Benatti of the Laboratory of Scanning Electron Microscopy of the Instituto de Botânica, São Paulo, Brazil, for the technical support and access to the scanning electron microscope. Also, the authors are grateful to the Coordination of the Rodoanel Mário Covas for the collection of material and its additon to the collections of the Herbarium SP. Funding This work was supported by the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) [grant numbers 2015/15359-0, 2016/ 24015-5, 2017/50341-0], Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), 18 C. F. P. da Luz et al. under the fellowship of ‘Productivity in research’ to Cynthia F. P. da Luz [grant number 304271/20195], and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for the post-doctoral grant awarded to Ana R. G. Simões (2015-208) [grant number 88881.067993/2014-01]. Disclosure Statement No potential conflict of interest was reported by the authors. Specimens Investigated Section 1. Dematostemon Griseb Dioscorea tauriglossum R. Knuth. Brazil: São Paulo, Iguape, Reserva Biológica Juréia-Itatins, trilha do Imperador, 16 May 1991, M. Kirizawa et al. 2520 (SP 248719); Brazil: São Paulo, root tuber collect in São Bernardo de Campo and cultivated at Instituto de Botânica, 20 February 2015, M. Kirizawa & J.H. Silva 3815 (SP 472163); Brazil: São Paulo, São Bernardo de Campo, Rodoanel Mário Covas, Southern stretche, 14 May 2008, J.H. Silva & M. Kirizawa 11 (SP 472166). Section 4. Sphaerantha Uline Dioscorea multiflora Mart. ex Griseb. Brazil: São Paulo, Novo Horizonte, mata Serrinha, Fazenda Rio Morto, 20 July 1994, R.R. Rodrigues et al. 22 (SP 289468); Brazil: São Paulo, Bananal, SP 247 highway back road, km 25, Alto da Serra, 7 April 2012, M. Kirizawa 3763 (SP 444406); Brazil: São Paulo, São Luiz do Paraitinga, Corredor Turístico São Sebastião, 11 November 2015, M. Kirizawa 3829 (SP 475315). Section 6. Chondrocarpa Uline Dioscorea altissima Lam. Brazil: São Paulo, Guarulhos, Rodoanel Mário Covas, Northern stretche, lote 6, OAE 604, R25, 25 March 2015, T. Shirasuna 3935 (SP 487481); Brazil: São Paulo, Arujá, Rodoanel Mário Covas, northern stretch, near to the Golf Club, 1 February 2013, M. Kirizawa & S. Aragaki 3766 (SP 446562); Brazil: São Paulo, Ubatuba, Picinguaba, trilha do Palmital, 14 January 1991, F.C.P. Garcia et al. 625 (SP 304317). Section 11. Monadelpha Uline Dioscorea monadelpha (Kunth) Griseb. Brazil: São Paulo, Guarulhos, Rodoanel Mário Covas, northern stretch, lote 6, 18 March 2015, R.T. Shirasuna 3910 (SP 487480); Brazil: São Paulo, Mauá, Rodoanel Mário Covas, southern stretch, lote 1, 23 April 2008, R.T. Shirasuna & E.A. Lopes 1215 (SP 426529); Brazil: São Paulo, root tuber collect in Mauá and cultivated at Instituto de Botânica, 15 December 2014, J.H. Silva & M. Kirizawa 13 (SP 472164). Section 43. Cryptantha Uline Dioscorea laxiflora Mart. ex Griseb. Brazil: São Paulo, Cananéia, Parque Estadual da Ilha do Cardoso, forest near the Tajuba hill, 17 April 1985, M. Kirizawa et al. 1446 (SP 202296); Brazil: São Paulo, Iguape, Reserva Biológica Juréia-Itatins, 16 May 1991, M. Kirizawa et al. 2510 (SP 248709); Brazil: São Paulo, Caraguatatuba, near to the beach, 30 April 1966, A.S. Grotta 342 (SP 287969). Dioscorea olfersiana Klotzsch ex Griseb. Brazil: São Paulo, Guarulhos, Rodoanel Mário Covas, northern stretch, lote 4, 8 April 2015, R.T. Shirasuna 3851 (SP 474884); Brazil: São Paulo, Ibiúna, Morro Grande district, Rio Una, c. 11 km of SP-250 highway, km 63, 28 December 1997, O. Yano & T. Yano 25144 (SP 318370); Brazil: Paraná, Curiúva, Fazenda São José da Boa Vista, Rio Barra Grande, 9 March 1999, E.M. Francisco s.n. (SP 377050). Section 44. Strutantha Uline Dioscorea amaranthoides C. Presl. Brazil: São Paulo, Caieiras, 2 March 1945, W. Hoehne s.n. (SP 49534); Brazil: São Paulo, Guarulhos, Rodoanel Mário Covas, Northern stretche, 26 March 2013, R.T. Shirasuna 3041 (SP 446580); Brazil: São Paulo, Mogi-Guaçu, Martinho Prado, Reserva Biológica da Fazenda Campininha, 27 January 1981, M. Sugiyama & W. Mantovani 99 (SP 168574). Section 47. Lasiogyne Uline Dioscorea dodecaneura Vell. Brazil: São Paulo, Iporanga, trilha do Carmo, 25 April 1995, M. Kirizawa et al. 3054 (SP 276042); Brazil: São Paulo, Piracicaba, mata da Pedreira, ESALQ/USP, 30 March 1988, E.L. Catharino 1205 (SP 298281); Brazil: São Paulo, Rodoanel Mário Covas, southern stretch, 2 April 2008, J.H. Silva & M. Kirizawa 12 (SP 246876). Section 48. Periandrium Uline Dioscorea trilinguis Griseb. Brazil: São Paulo, São Bernardo do Campo, Rodoanel Mário Covas, southern stretch, lote 3, 26 May 2008, R.T. Shirasuna 1354 (SP 426533); Brazil: São Paulo, Santo André, Reserva Biológica do Alto da Serra de Paranapiacaba, 6 July 1985, M. Kirizawa 1495 (SP 212491); Brazil: São Paulo, Jabaquara, 3 April 1946, A. Gehrt s.n. (SP 53704). ORCID Cynthia Fernandes Pinto Da Luz http://orcid.org/00000001-7908-155X http://orcid.org/0000-0001-7267Ana Rita G. Simões 8353 References Alzer FC. 2017. Palinotaxonomia de representantes de Dioscorea subgenera Helmia e Dioscorea subgênero Dioscorea (Dioscoreaceae) ocorrentes no neotrópico. PhD Thesis, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil. Barroso GM, Sucre D, Guimarães EF, Carvalho LF, Valente MC, Silva JD, Silva JB, Rosenthal FR, Barbosa CM, Barth OM, Barbosa AF. 1974. Flora da Guanabara - Família Dioscoreaceae. Sellowia 25: 9–256. Caddick LR, Furness CA, Stobart KL, Rudall PJ. 1998. Microsporogenesis and pollen morphology in Dioscoreales and allied taxa. Grana 37: 321–336. doi:10.1080/ 00173139809362687. Caddick LR, Rudall PJ, Wilkin P. 2000. Floral morphology and development in Dioscoreales. Feddes Repertorium 111: 189– 230. doi:10.1002/fedr.20001110313. Caddick LR, Wilkin P, Rudall PJ, Hedderson TAJ, Chase MW. 2002. Yams reclassified: A recircumscription of Dioscoreaceae and Dioscoreales. Taxon 51: 103–114. doi:10.2307/ 1554967. Cain SA, Cain DE. 1948. Size-frequency characteristic of Pinus echinata. Botanical Gazette 110: 325–330. Corrêa AMS. 1996. Flora polinica da Reserva do Parque Estadual das Fontes do Ipiranga (Sao Paulo, Brasil). Familia 196 Dioscoreaceae. Hoehnea 23: 103–106. Couto RS, Martins AC, Bolson M, Lopes RC, Smidt EC, Braga JMA. 2018. Time calibrated tree of dioscorea (dioscoreaceae) indicates four origins of yams in the neotropics since the eocene. Botanical Journal of the Linnean Society 188: 144– 160. Couto RS, Tenorio V, Alzer FC, Lopes RC, Vieira RC, Mendonça CBF, Gonçalves-Esteves V, Braga JMA. 2014. Taxonomic Revision of the Dioscorea campestris Species Pollen morphology of Dioscorea Assemblage (Dioscoreaceae). Systematic Botany 39: 1056– 1069. doi:10.1600/036364414X683895. Ejsmond MJ, Ejsmond A, Banasiak L, Karpin´ska-kołaczek M, Kozłowski J, Kołaczek P. 2015. Large pollen at high temperature: An adaptation to increased competition on the stigma? Plant Ecology 216: 1407–1417. doi:10.1007/s11258-015-0519-z. Erdtman G. 1952. Pollen morphology and plant taxonomy: Angiosperms. Stockholm: Almqvist & Wiksell. Erdtman G. 1960. The acetolysis method. A revised description. Svensk Botanisk Tidskrift 54: 561–564. Govaerts R, Wilkin P, Saunders RMK. 2007. World checklist of Dioscoreales: Yams and their allies. Kew: Royal Botanic Gardens. Gupta HP, Sharma C. 1986. Pollen Flora of North-West Himalaya. Lucknow: Indian Association of Palynostratigraphers. Halbritter H, Grímsson SUF, Weber M, Zetter R, Hesse M, Buchner R, Svojtka M, Frosch-Radivo A. 2018. Illustrated pollen terminology. Cham: Springer International Publishing. Hesse M, Halbritter H, Weber M, Buchner R, Frosch-Radivo A, Ulrich S. 2009. Pollen terminology – An illustrated handbook. Vienna: Springer. Heusser CJ. 1971. Pollen and spores of Chile: Modern types of the Pteridophyta, Gymnospermae, and Angiospermae. Tucson, AZ: The University of Arizona Press. Huang TC. 1970. Pollen grains of Formosan plants. Taiwania 15: 73–179. Ikuse M. 1956. Pollen grains of Japan. Tokyo: Hirokawa. Jones M, Newelll LC. 1942. Size variability and identification of grass pollen. Journal of the American Society of Agronomy 40: 136–143. doi:10.2134/agronj1948.00021962004000020004x. Kirizawa M, Xifreda CC. 2009. Dioscoreaceae. In: Stehmann JR, Forzza RF, Salino A, Sobral M, Costa DP, Kamino LHY, eds. Plantas da Floresta Atlântica, 238–239. Rio de Janeiro: Jardim Botânico do Rio de Janeiro. Kirizawa M, Xifreda CC. 2011. Dioscoreaceae. In: Wanderley MGL, Shepherd GJ, Martins SE, Estrada TEMD, Romanini RP, Koch I, Pirani JR, Melhem TS, Giulietti-Harley AM, Kinoshita LS, Magenta MAG, Wagner HML, Barros F, Lohmann LG, Amara MCE, Cordeiro I, Aragaki S, Bianchini RS, Esteves GL, eds. Checklist das Spermatophyta do Estado de São Paulo, 253–254. São Paulo: Biota Neotropica 11 (supl.). Kirizawa M, Xifreda CC, Couto R, Araújo D. 2015. Dioscoreaceae. In: Lista de espécies da flora do Brasil. Jardim Botânico do Rio de Janeiro. http://floradobrasil.jbrj.gov.br/jabot/florado brasil/FB7372; accessed 15 November 2019. Knuth RGP. 1924. Dioscoreaceae. In: Engler HGA, ed. Das Pflanzenreich 4, 43: 1–387. Martin ARH. 1959. South African palynological studies I. Statistical and morphological variation in the pollen of the South African species of Podocarpus. Grana Palynologica 2: 40–68. doi:10.1080/00173135909431977. McCune B, Melfford MJ. 1999. PC-ORD. Multivariate analysis of ecological data. Gleneden Beach, OR: MjM Software Design. Miège J. 1965. L’appui de la palynologie dans la distinction des espèces africaines de Dioscorea. Webbia 19: 841–845. Punt W, Hoen PP, Blackmore S, Nilsson S, Le Thomas A. 2007. Glossary of pollen and spore terminology. Review of Paleobotany and Palynology 143: 1–81. doi:10.1016/j.revpalbo.2006.06.008. Raynal A, Raynal J. 1971. Une technique de preparation des grains de pollen fragiles. Adansonia 11: 77–79. Roubik DW, Moreno PJE. 1991. Pollen and spores of Barro Colorado Island. Monographs in Systematic Botany from the Missouri Botanical Garden 36: 83–84. Salgado-Labouriau ML, Vanzolini PE, Melhem TS. 1965. Variation of polar axes and equatorial diameters in pollen grains of two species of Cassia. Grana 6: 166–176. 19 Schols P, Furness CA, Merckx V, Wilkin P, Smets E. 2005a. Pollen Evolution in Yams (Dioscorea: Dioscoreaceae). Systematic Botany 30: 750–758. doi:10.1600/036364405775097743. Schols P, Furness CA, Wilkin P, Huysmans S, Smets E. 2001. Morphology of pollen and orbicules in some Dioscorea species and its systematic implications. Botanical Journal of the Linnean Society 136: 295–311. doi:10.1111/boj.2001.136.issue3. Schols P, Furness CA, Wilkin P, Smets E, Cielen V, Huysmans S. 2003. Pollen morphology of Dioscorea (Dioscoreaceae) and its relation to systematics. Botanical Journal of the Linnean Society 143: 375–390. doi:10.1111/j.1095-8339.2003.00227. x. Schols P, Wilkin P, Furness CA, Huysmans S, Smets E. 2005b. Comparative pollen development in Dioscoreales. International Journal of Plant Sciences 166: 909–924. doi:10.1086/ 449316. Sharma M. 1967. Pollen morphology of Indian monocotyledons. J. Palynology (special volume). Lucknow: Palynological Society of India. Sheikh N, Kumar Y. 2018. Palynological Implication to the Systematic of the genus Dioscorea in Meghalaya, North East India. International Journal of Pure and Applied Bioscience 6: 94–100. doi:10.18782/2320-7051.6049. Shepherd GJ. 1996. Fitopac 1: Manual do usuário. Campinas: Universidade Estadual de Campinas. Silva JH, Barbuglio-Santiago LO, Kirizawa M. in press. Dioscoreceae do Rodoanel Mário Covas do Estado de São Paulo, Brasil: Diversidade florística. Hoehnea. Su P. 1987. Pollen morphology of Dioscorea in China. Acta Phytotaxonomica Sinica 25: 257–365. Thiers B. continuously updated. Index Herbariorum: A global directory of public herbaria and associated staff. New York Botanical Garden’s Virtual Herbarium. http://sweetgum.nybg. org/science/ih/. Vieira S. 2008. Introdução à Bioestatística. Rio de Janeiro: Elsevier. Viruel J, Segarra-Moragues JG, Raz L, Forest F, Wilkin P, Sanmartín I, Pilar Catalán P. 2016. Late Cretaceous–Early Eocene origin of yams (Dioscorea, Dioscoreaceae) in the Laurasian Palaearctic and their subsequent Oligocene–Miocene diversification. Journal of Biogeography 43: 750–762. doi:10.1111/jbi.12678. Wilkin P, Davis KK, Schols P, Furness CA. 2002. Dioscorea ridleyi and D. palawana: New systematic data on two rare South-East Asian yam species. Kew Bulletin 57: 885–900. doi:10.2307/4115719. Wilkin P, Muasya AM, Banks H, Furness CA, Vollesen K, Weber O, Demissew S. 2009. A new species of yam from Kenya, Dioscorea kituiensis: Pollen morphology, conservation status, and speciation. Systematic Botany 34: 652–659. doi:10.1600/036364409790139646. Xifreda CC. 1982. Estudios en Dioscoreáceas I. Boletín de La Sociedad Argentina de Botánica 20: 317–327. Xifreda CC. 2000. Evaluation of pollen and vegetative characters in the systematics of South American species of Dioscorea (Dioscoreaceae). In: Wilson KL, Morrison DA, eds. Proceedings of the 2nd International Monocot Conference: Systematics and evolution, 488–496. Melbourne: CSIRO. Xifreda CC, Kirizawa M. 2003. Observaciones morfológicas, nomenclaturales y taxonómicas en dioscórea monadelpha yD. Subhastata (Dioscoreaceae). Darwiniana 41: 207–215. Zavada MS. 1983. Comparative morphology of monocot pollen and evolutionary trends of apertures and wall structures. The Botanical Review 49: 331–379. doi:10.1007/BF02861086.