Oncophoraceae (Bryophyta): a palynological treatment of species occurring in the Americas

: Oncophoraceae are acrocarpous mosses that predominantly grow as tufts or cushions and especially occur on rocks and soil. The recognition of Oncophoraceae as a distinct family, as well as its generic circumscription, is not consensus among authors, and the pursuit for new information to improve its characterization is incessant. The present work aims to characterize the spore morphology and ultrastructure of 19 species (eight genera) occurring in the Americas and to evaluate the relevance of palynological data to circumscribe species, contributing to support other palynological studies. Observations were performed under Light and Electron (Scanning and Transmission) Microscopes. A Cluster Analysis was performed in order to evaluate the meaning of the palynological data, especially concerning the establishment of the species circumscription. Spores are monads, small to medium sized (10.40 to 44.20 μm), radially symmetric, subcircular in amb, heteropolar or apolar; the surface is ornamented by granules, gemmae and bacula. Anisomorphic spores were observed in eight studied species and are reported herein for the fi rst time. The Cluster Analysis shows two groups with low similarity, which primarily differ by the polarity of the spores. The circumscription of Kiaeria and Cynodontium is corroborated by palynological


INTRODUCTION
Oncopho raceae M. Stech are acrocarpous mosses, small to medium sized, growing in tufts or cushions, rarely pendant, on rocks or soil, and less frequently on trees.The stems are short, simple or sparingly branched; the leaves are lanceolate to narrow-lanceolate and oblong at the base.The capsules are immersed to exserted, ovoid to pyriform; the peristome is simple, having 16 teeth (Frahm 2002, Frey & Stech 2009, Goffi net et al. 2008, Gradstein et al. 2001, Stech & Frey 2008).
Since the 19 th and 20 th centuries, these genera were mostly treated in Dicranaceae (Schimper 1856) or Rhabdoweisiaceae (Limpricht 1904) in different circumscriptions, including variations in levels of subfamilies or tribes (Brotherus 1924, Crosby et al. 1999, Fleischer 1900, Vitt 1984).Stech (1999a, b) highlighted the need for taxonomic revision of the group, which was sought by La Farge et al. (2002), Ochyra et al. (2003), Tsubota et al. (2003), Hedderson et al. (2004), Stech & Frey (2008) and Zander (2008).But the relationship between these species has not yet been completely resolved (Cox et al. 2010, Stech et al. 2012).Increasing the number of morphological information of Oncophoraceae species is very important to support the taxonomy, especially considering palynological data, which are still scarce for the group.
The palynological study proposed herein aims to describe the spore morphology and ultrastructure of 19 Oncophoraceae species occurring in the Americas, aiming to broaden the morphological data employed for their characterization, as well as to provide information to support palynological studies from different occurrences.

MATERIALS AND METHODS
We employed herbarium material from the following collections to perform this study: the Canadian Museum of Nature Herbarium (CANM), the Maria Eneyda P. Kauffmann Fidalgo Herbarium (SP), the Universidade de Brasília Herbarium (UB), and the Museu Nacional do Rio de Janeiro Herbarium (R).The acronyms are in accordance to Thiers (2020).
We studied species that occur in the Americas, but some of them are also present in other continents.In order to examine as many specimens as possible, all available exsiccates that presented mature capsules were examined, even if they were from another continent.The available material which had mature capsules were selected, totalling 19 species, namely: Arctoa fulvella (Dicks.tetroxide; dehydrated in a graded ethanol series, and then dried out in a Critical Point dryer.Spores were subsequently dispersed upon stubs with double-sided carbon tape, and covered with a 20nm gold layer to be observed. For observations under TEM, the capsules were fixed in glutaraldehyde and post-fixed in osmium tetroxide, washed in buffer solution and dehydrated in a graded ethanol series.They were encased in Spurr resin (Spurr 1969) and heated to 70°C for 48h; the resin blocks were then sectioned and mounted on TEM copper mesh.The material was contrasted with uranyl acetate and lead citrate (Reynolds 1963) and then observed.
The largest diameter measures were obtained using a micrometric ocular coupled to a Light Microscope.One sample material was indicated as the standard for each species (indicated by * on Tables), and the other sample materials as comparison.For all isosporic taxa, three slides were prepared and 50 spores were taken at random for larger diameter measurements.For all anisosporous species, the largest diameter of 100 spores was measured randomly for both standard and comparison material.For heteropolar spores, in order to take polar (P) and equatorial (E) diameter measurements, 30 spores were observed in equatorial view.
Arithmetic mean (X), standard deviation (S), standard error (Sx), coefficient of variability (CV%), and 95% confidence interval (CI) are presented, as well as the minimum and maximum spore size (Xmin-Xmax) of each analysed material.
The data referring to the larger diameter, polar and equatorial diameter values did not meet the normality assumptions, and were therefore statistically analysed using the Kruskal Wallis test.Next, a posteriori Dunnett test was applied to identify the different treatments.The statistical analysis was developed in Past.2.17c (Hammer et al. 2001).
By virtue of the slight thickness of perine and exine, these two layers were measured together, configuring the sclerine (Luizi-Ponzo & Barth 1998, 1999).Sclerine and intine were measured from 10 random spores, prepared according to the Wodehouse (1935) method (and observing the adjustment by Luizi-Ponzo & Melhem 2006b), and arithmetic means were obtained.
Line graphs are presented to evaluate the spore size distribution, in which the values of the larger spore diameters were included in frequency classes.For anisosporous species, spores within the range of the first peak class of the line graph were referred to as "smallest spores", and spores within the range of the second peak are called "largest spores" (Rodrigues & Luizi-Ponzo 2015).
A binary matrix was elaborated to perform the Cluster Analysis with the palynological data of the analysed species, namely, spore size, polarity, presence of anisospory, sporoderm thickness, type of ornamentation and aperture area.The data were submitted to Cluster Analysis on Past ver.2.17c software (Hammer et al. 2001), and using the Jaccard similarity index to verify the degree of similarity between the species.

RESULTS
The studied taxa have small to medium sized spores (Tables I, II), isomorphic or anisomorphic, in monads, with radial symmetry, subcircular amb, heteropolar or apolar (Tables III, IV); apertural region differentiated or not, subcircular, and may present distinct ornamentation from Xmin-Xmax: minimum and maximum values of spores size diameter, X ± Sx: mean and standard error, S: standard deviation, CV%: coefficient of variation, IC 95%: confidence interval.
the rest of the surface.Sporoderm is formed by intine, exine and perine, measuring between 1.16 µm and 2.93 µm (Tables V, VI).It was necessary to observe the spores under SEM for detailed description of the ornamentation due to the small size of the spores and discreet ornamentation processes.
A Cluster Analysis (Fig. 4) using the palynological characteristics (Table VII) shows a cophenetic index of 0.9634, demonstrating that the characteristics used for analysis are consistent, although the variation in spore size is large (Fig. 5).It was possible to group the 19 studied species into two large groups (Group A and Group B), and then into six subgroups (B1 to B6).
Group A: formed by the Kiaeria species: K. falcata, K. glacialis, and K. starkei, species which present apolar spores.
The spores are small in size (Table II) with unimodal distribution (Fig. 6a-c).The sporoderm surface is ornamented with granula, which may be grouped or overlapped, with different sizes and distributions.The exine is fully covered by the perine or it has small exposed areas (Figs. 1i,1j,.
Group B: formed by the other studied species, all of them have heteropolar spores.This group was divided into six subgroups.
Subgroup B1: formed exclusively by the Oreoweisia laxiretis, species which has spores without a defined apertural area.The spores are small in size (Table II) with unimodal size distribution frequency (Fig. 6d) and the sporoderm surface is heavily ornamented with gemma, which can be uniform or varied in size and exhibit overlap (Figs. 1k,2a,2b).
Subgroup B2: formed by Cynodontium gracilescens, C. polycarpon, C. strumiferum, C. strumulosum, and C. tenellum, which present anisomorphic spores and sporoderm surface with different kinds of granula processes.The spores are small to medium in size (Table I), with bimodal spore size distribution (Fig. 6ei).The sporoderm surface is ornamented with bacula, which may be single or united, gemma and granula.An apertural subcircular area is present, having a distinct ornamentation from the remaining surface of the sporoderm .
Subgroup B5: formed exclusively by Rhabdoweisia fugax, it shows isomorphic spores and sporoderm surface with ornamentation processes other than granula.The spores are small in size (Table II) with unimodal spore size distribution (Fig. 6q).The sporoderm surface is ornamented by bacula which can be single or grouped, and exine shows exposed areas (Figs. 3n,3o).
Subgroup B6: includes Arctoa fulvella and Oreoweisia brasiliensis, which present isomorphic spores and granulate sporoderm surface.The spores are small in size (Table I) with unimodal spore size distribution (Fig. 6r-s).The sporoderm surface is granulate, the granula can be grouped or overlapped with different sizes and distributions, and there are gemma in the apertural area.The exine is fully covered by the perine or shows small exposed areas (Figs. 1a,1b,2d,2e,2m   One species in Arctoa presented isomorphic spores (A.fulvella), and the other anisomorphic spores (A.hyperborea).Frisvoll (1978) says that A. fulvella presents spores with larger diameter between 14-24µm, and Ochyra & Buck (2003) cite spores as globose, with a rough surface and larger diameter between 18-22 µm.Newmaster (2017) says that A. fulvella spores have diameter between 16-28 µm and A. hyperborea about 16-30 µm; these measurements are close to those found in this study.Newmaster (2017) indicates Cynodontium spores measuring about 10µm to 25µm, and describes them as smooth to baculate, while Oncophorus spores are described by him as gently rough, measuring about 14µm to 25µm.These measurements are compatible with those found herein to the "smallest spores" of these genera; however, Newmaster (2017) does not mention the occurrence of anisospory.
Our results demonstrate that the species of Cynodontium studied are included in a single morphological type of spores, supporting the taxonomical interpetation of these species.However, species of Oncophorus were grouped in the same palynological type, but together with a species of Arctoa (A. hyperborea), showing the morphological complexity of these species.
Luizi-Ponzo & Barth (1999) described Oreoweisia brasiliensis spores.They indicated the measurements of the spore diameter as being about 20.80 µm to 30.40 µm, while the mean was 24.10 µm ± 0.40 µm.The spore size range is larger than that found in this study (18.20 µm -28.60 µm), but the mean is close (24.20 µm ± 0.38 µm), while the granulate surface and the apertural area fit in both studies.
Tan & Schofield (1980), Schofield (2017) and Weber (2017) reported the spores of Dicranoweisia and Oreas martitana employing a different terminology, but similarities are observed with the specimens examined here.However, for O. martiana, Weber (2017) cites spores of about 16 µm; this is quite different from those spores reported herein, as we observed a higher amplitude and different mean size.
Newmaster (2017) characterized the spores of Kiaeria as spherical, measuring about 14µm  (2022) 94(1) e20201508 12 | 20   and 24µm; these values are near to those observed herein.All Kiaeria species studied were grouped in the same morphological type of spores, corroborating the their generic circumscription.
While studying Amphidiaceae spores, Passarella & Luizi-Ponzo (2019) considered them to be isomorphic, small in size and with a strong heteropolar condition, in which the distal faces of the spores are perforated, and the proximal faces exhibited an apertural area surrouded by gemma and rugulae connected.Our results demonstrated that these conditions are not found in the spores of Oncophoraceae, favoring the separation of families, as proposed by Frey & Stech (2009).

CONCLUSION
Oncophoraceae species present small to medium spores with radial symmetry, subcircular amb, they are heteropolar or apolar.The sporoderm stratifi cation includes perine, exine and intine.Eight species studied, representing three genera, present anisomorphic spores; this was not reported before, according to the studied literature.
The small size of the spores indicates the importance of SEM observations to refi ne the description of sporoderm of these species.
Kiaeria species: K. falcata, K. glacialis and K. starkei may be defi ned by the granulate surface of the spores; and the anisomorphic baculate spores of Cynodontium characterize the species of this genus.
Spore size, ornamentation and sporoderm stratification measurements vary between Oncophoraceae species, which allows us to say that the family is euripalynous.Despite the great morphological variability observed in the spores of the species of Oncophoraceae studied, their distinction from Amphidiaceae is here corroborated.

Figure 6 .
Figure 6.Line graphs representing the spore size frequency distribution of the species.

Table I .
Morphometric data of anisomorphic spores (measurements in micrometers, * standard material, the other, comparison ones).

Table II .
Morphometric data of isomorphic spores (measurements in micrometers, * standard material, the other, comparison ones). ).
Xmin-Xmax: minimum and maximum values of spores size diameter, X ± Sx: mean and standard error, S: standard deviation, CV%: coefficient of variation, IC 95%: confidence interval.

Table III .
Mean values of equatorial and polar diameters of heteropolar anisomorphic spores (in micrometers).

Table IV .
Mean values of equatorial and polar diameters of heteropolar isomorphic spores (in micrometers).

Table V .
Mean values of the sporoderm strata thickness and aperture diameter of anisomorphic spores (in micrometers).

Table VI .
Mean values of the sporoderm strata thickness and aperture diameter of isomorphic spores (in micrometers).