Pollen morphology of the African Sclerosperma (Arecaceae)

Abstract Three currently accepted Sclerosperma species appear to produce four different pollen morphologies. Sclerosperma mannii and S. walkeri pollen share the same distinct reticulate sculpture, but S. profizianum produces three different pollen types (microreticulate, fossulate, and perforate). The pollen morphology suggests that S. mannii and S. walkeri are sister taxa of the same intrageneric lineage. The pollen diversity observed in S. profizianum suggests (a) this taxon is unique regarding its pollen diversity despite being a non-heterostylous plant or (b) that circumscription of S. profizianum as a species may well be in the need of redefinition.

; Van Valkenburg et al. 2008). The first account of Sclerosperma pollen by Erdtman and Sing (1957) documented its unique morphology within the palm family (triangular, triporate, reticulate), features often discussed by M.M. Harley in the years 1991 to 2008 through her extensive work with colleagues on the pollen morphology of Arecaceae (Harley & Hall 1991;Harley 1996Harley , 1999Harley , 2004Harley & Baker 2001;Harley & Dransfield 2003;Dransfield et al. 2008). Despite the number of publications containing pollen descriptions and micrographs of the extant genus, Sclerosperma, a detailed characterisation of the pollen morphology of the three species was needed. Also, the taxonomy of this genus was only recently revised (Van Valkenburg et al. 2008), and showed that previously published pollen material often originated in misidentified specimens. SEM). The classification and author names of extant species follow WCSP (2018). Classification above genus level follows Dransfield et al. (2008) and APG IV (2016). Herbarium materials were assigned to extant species according to Van Valkenburg et al. (2008). Pollen grains of each taxon are described individually. Pollen grains of the three Sclerosperma species are also compared with each other in Table III.
Note regarding the following descriptions. -According to Halbritter et al. (2018) an ulcus (pl. ulci) is a more or less circular aperture situated distally on the pollen. Ulci are confined to gymnosperms, magnoliid and monocot angiosperm taxa. Also, according to Halbritter et al. (2018) a porus (pl. pori) is a more or less circular aperture located at the equator or regularly spread over the pollen grain. Pori are confined to dicot angiosperm taxa. Still, in all the literature regarding Sclerosperma by M.M. Harley from the years 1991 to 2008 (Harley & Hall 1991;Harley 1996Harley , 1999Harley , 2004Harley & Baker 2001;Harley & Dransfield 2003;Dransfield et al. 2008) the pollen of this genus was described as porate. In order to avoid confusion the apertures of Sclerosperma are here also termed pori and the pollen is regarded as triporate and not triulcerate despite the distal position of the apertures.  Table III) Description. -Pollen, monad, heteropolar, polar axis/equatorial diameter (P/E) ratio oblate, outline straight-triangular to slightly concave-triangular in polar view, bean-shaped in equatorial view (convex distal face versus concave proximal face); equatorial diameter 32-38 µm in LM, 27-34 µm in SEM, polar axis 9-15 µm in LM; triporate, pori positioned subapically on the distal polar face, pori elliptic, 4.5-6.0 µm in diameter, pori equipped with opercula; exine 1.7-2.5 µm thick in LM, nexine thinner than sexine; pollen wall semitectate; sculpture reticulate in LM, reticulate to perforate in SEM; distal face reticulate with broad muri and elliptic to triangular to polygonal lumina, 18-25 lumina per 100 µm 2 at central distal face, 0-6 nanogemmae free-standing columellae per lumina (SEM); proximal face reticulate to perforate, lumina/perforations elliptic to triangular to polygonal, 0-6 nanogemmae free-standing columellae per lumina; central polar areas and interapertural areas reticulate, sculpture becoming microreticulate to perforate towards apices; opercula with nanoverrucate to granulate sublayer and distinct microreticulate supra-layer (SEM).
Remarks. -The first LM and SEM micrographs showing pollen of this taxon are by Harley and Hall (1991, plate 4, figures 32 [SEM] and 33 [LM]). The same SEM micrograph is shown in Harley (1999, plate 1, figure 12), and again along with two additional LM micrographs and two attached grains under SEM in Harley (1996, plate 16, figures C and F [SEM], and G and H [LM]). These are all repeated in Harley and Baker (2001, figures 77 and 82 [SEM], 80 and 81 [LM]). A new SEM detail is provided in Harley and Dransfield (2003, figure 11). In total, five or six grains were illustrated using either LM or SEM micrographs, but no TEM micrograph has been presented thus far. All the pollen grains of this taxon illustrated by Harley (1996), Harley and Baker (2001) and Harley and Dransfield (2003) originate from the same herbarium sample (Tuley, s.n. [K]). All other previously published micrographs showing alleged pollen of Sclerosperma mannii originated from misidentified specimens (see Table IV). The LM and SEM based pollen morphology of the Tuley s.n. (K) sample is similar to that now observed in the van der Burgt 1958 (K) sample. The only difference is that the free-standing columellae are more prominent and more frequent in the specimen collected by Tuley versus that collected by van der Burgt, a feature comparable to what is observed in the S. walkeri pollen from the Leonard 1614 (BR) sample (compare figure 11 in Harley and Dransfield [2003] Table III) Description. -Pollen, monad, heteropolar, P/E ratio oblate, outline straight-triangular to slightly concave-triangular in polar view, bean-shaped in equatorial view (convex distal face versus concave proximal face); equatorial diameter 35-40 µm in LM, 29-35 µm in SEM, polar axis 10-14 µm in LM; triporate, pori positioned sub-apically on the distal polar face, pori elliptic, 5.0-8.5 µm in diameter, pori equipped with opercula; exine 1.7-2.5 µm thick in LM, 2.1-2.5 µm thick in SEM, nexine thinner than sexine, nexine 0.5-0.8 µm thick        figure 6F), and an additional SEM micrograph is provided in Dransfield et al. (2008, p. 391, figure b). All previously illustrated pollen grains that are here referred to S. profizianum (pollen Type A) were formerly assigned to S. mannii (see Table IV).  (2003), Harley (2004) and Dransfield et al. (2008), but pollen Type C (Table IV). This suggests that S. profi- Pollen morphology of Sclerosperma     Note: Species affiliation according to Van Valkenburg et al. (2008). The drawings presented by Erdtman and Singh (1957) are not included in this table since they are not affiliated to any of the three Sclerosperma samples used in their study.
Remarks. -The first SEM micrographs showing pollen of this taxon are by Harley (1996, pl. 16, figures D and E). The same two SEM micrograph are repeated in Harley and Baker (2001, figures 78 and 79). A single TEM showing the aperture region and operculum is presented in Harley and Dransfield (2003, figure 18). All previously illustrated pollen grains belonging to this taxon (pollen Type C) were formerly assigned to Sclerosperma mannii (see Table IV).

Discussion
Differentiating Sclerosperma pollen Based on the pollen morphology of Sclerosperma presented herein it is clear that there are at least four different pollen morphologies produced by the three extant taxa. Sclerosperma mannii and S. walkeri share similar pollen morphology, and are difficult to distinguish from each other in both LM and SEM. Sclerosperma profizianum produced three different pollen morphologies (Types A, B, and C), distinguishable from each other and from S. mannii and S. walkeri, particularly in SEM.
Using LM only, Sclerosperma pollen can be divided into reticulate (including S. manni, S. profizianum Type A, and S. walkeri) and non-reticulate (including S. profizianum Type B and C; Table III). The reticulate pollen are further divided into coarsely reticulate (including S. mannii and S. walkeri) versus finely reticulate (S. profizianum Type A; compare Figure 1A and 1Z with Figure 1F). Our measurements indicate the coarsely reticulate pollen of S. mannii and S. walkeri can be set apart using the length of their polar axis, which is longer in the pollen of S. walkeri (15-19 µm) than in S. mannii (9-15 µm). The non-reticulate Sclerosperma pollen grains are distinguishable by having rugulate (S. profizianum Type B) or scabrate (S. profizianum Type C) sculpture. The rugulate S. profizianum Type B pollen is usually smaller than that of S. profizianum Type C (Table III), and the outline of the pollen in polar view is more convex-triangular in S. profizianum Type C versus concave-triangular in S. profizianum Type B.
Applying additional SEM, there are a number of details separating the three different Sclerosperma pro-Pollen morphology of Sclerosperma fizianum pollen types from each other and from the pollen of S. mannii and S. walkeri. The magnification obtained using SEM shows that S. mannii and S. walkeri pollen is more or less identical (Table III). The only noticeable sculpture difference, so far, is that the free-standing columellae in S. walkeri pollen are more frequent and conspicuous than in S. mannii pollen (compare Figure 2E and 2F with Figure 6E and 6F). Still, the S. mannii and S. walkeri pollen are easily distinguished from the three S. profizianum pollen types. The S. mannii and S. walkeri pollen is reticulate with 16-25 lumina per 100 µm 2 at the central distal face versus microreticulate with 30-35 lumina per 100 µm 2 in S. profizianum Type A (compare Figures 2E and 6E with Figure 3E). The S. profizianum Type C pollen is perforate and rugulate/verrucate with 45-55 perforations per 100 µm 2 at the central polar face, and the S. profizianum Type B pollen is fossulate and rugulate/verrucate with the perforations hidden in the fossulae (compare Figure 5E with Figure 4E; Table III). Furthermore, the S. profizianum Type A pollen has regularly distributed elliptic to circular or slit-like lumina versus perforations aligned in sinuous rows in S. profizianum Type C versus tiny perforations aligned (hidden) within the fossulae in S. profizianum Type B pollen. Also, the operculum in S. profizianum Type A pollen has a clear reticulate supra-layer versus perforate in both S. profizianum Types B and C pollen (compare Figure 3C with Figures 4C  and 5C).

Pollen morphology and taxonomic resolution
The 'identical' pollen of Sclerosperma mannii and S. walkeri are from sites near the centre of distribution for the genus (see map 1 in Van Valkenburg et al. 2008). Pollen of S. profizianum Type C is from a disjunct population in Ghana and is also found in a more centrally located population in Republic of the Congo. The two other S. profizianum pollen Types, A and B, are from the southern edge of the distribution of that species/genus near the border between Democratic Republic of the Congo and Angola.
There are many examples of clades in which species can be clearly separated on the basis of plant body and reproductive parts, but the pollen produced by them are similar or identical to each other morphologically (termed stenopalynous taxa, see Halbritter et al. 2018). However, it is not common for a single species to produce two or more distinct pollen morphologies (in sculpture and/or size), unless the plants are heterostylous (for a list of such genera see table 1 in Ganders 1979). There are no hints in the literature that heterostyly occurs in Sclerosperma.
We did not discover more than a single pollen type from a particular or several anthers out of an individual Sclerosperma flower; the pollen morphology observed within an anther or anthers were distinct and consistent within an individual or between flowers from the same herbarium sample. The pollen morphology of S. profizianum Types B and C suggests that they are very close, and based on previous work by Harley on some of the same herbarium material [Profizi, 841 (K)] it is even possible that they were produced by the same plant. Still, M.M. Harley only figured a single pollen grain in SEM (Harley & Dransfield 2003;Harley 2004;Dransfield et al. 2008) and therefore a sampling error or contamination cannot be excluded. The material studied might also have been assigned to the wrong collector information. Disregarding all that, and assuming that the Types B and C pollen originate from the same plant or taxon it is clear that S. profizianum still seems to produce two undoubtedly different pollen types: (1) the microreticulate pollen Type A and (2) the fossulate/perforate Type B/C pollen.
The reason Sclerosperma profizianum seems to produce different pollen types is unclear currently, but we can think of two possible explanations for this situation. First, S. profizianum is a unique taxon that produces different pollen types without being heterostylous. This seems very unlikely, but cannot be excluded. Second, the currently accepted species definitions in Sclerosperma do not reflect its actual biological diversity. Sclerosperma profizianum may be composed of more than a single natural species, or at least it may be in the process of genetic diversification related to its disjunct distribution and marginal occurrences. This could explain observed variations in pollen morphology, including the intermediate sculpture features of S. profizianum pollen Type A (microreticulate versus reticulate in S. mannii/ S. walkeri, and perforate in S. profizianum Type B/C). Whatever the explanation, it cannot be resolved from the data presented here.

Conclusion and outlook
Combined LM and SEM analyses demonstrate that there are four different pollen morphologies produced by Sclerosperma. Sclerosperma mannii and S. walkeri share similar pollen morphologies, but S. profizianum produces three different pollen types. Despite the detailed pollen work presented here and all the available literature regarding Arecaceae taxonomy, pollen morphology and phylogeny (e.g. Dransfield et al. 2008), there is still much that needs to be studied. For Sclerosperma, it would be vital to explore intrageneric relationships and conduct a 112 F. Grímsson et al. molecular phylogenetic study using several representatives from each alleged species to see how they align in a phylogenetic tree. In this sense it would be interesting to see if the S. mannii samples group together and appear as sister taxon to the S. walkeri samples, and if all the S. profizianum fall into one branch or are not clearly resolved suggesting some sort of species differentiation. When conducting such a study it would be highly informative to analyse pollen morphology from the same plants used for the molecular phylogeny and plot the pollen on the resulting tree. In such a case the evolution of pollen morphology in Sclerosperma could be resolved. Also, the comprehensive LM and SEM based pollen morphology presented here will now allow for a better determination of fossil Sclerosperma pollen grains and the re-analyses of fossil material previously affiliated to this genus.