Biometrics of the sporangia and spores of the Parablechnum cordatum complex (Blechnaceae, Polypodiopsida)

Parablechnum is the most diverse genus of Blechnaceae (ca. 65 species), with a pantropical distribution and two centers of diversity, in America and in the Austro-Pacific region. The species are dimorphic, with often erect rhizomes and rhizomatic scales, 1-pinnate fronds, with truncate blade at base, conform apex and stalked pinnae. This group presents many taxonomic problems, needing more detailed studies to resolve these conflicts of separation between species. This work deals with the American complex of P. cordatum in which the species P. cordatum, P. schiedeanum, P. chilense, P. falciforme and Blechnum varians are included. A biometric analysis of sporangia and spores, important taxonomic structures in the distinction of ferns, has been carried out. The data were subjected to a one-way analysis of variance and a discriminant analysis. In addition, the spores were observed under a scanning electron microscope to study their ornamentation. Of the characters we have studied thickness of the arcus, number of cells in the arcus, number of cells in the hypostome and major equatorial diameter of the spore have statisticalParablechnum is the most diverse genus of Blechnaceae (ca. 65 species), with a pantropical distribution and two centers of diversity, in America and in the Austro-Pacific region. The species are dimorphic, with often erect rhizomes and rhizomatic scales, 1-pinnate fronds, with truncate blade at base, conform apex and stalked pinnae. This group presents many taxonomic problems, needing more detailed studies to resolve these conflicts of separation between species. This work deals with the American complex of P. cordatum in which the species P. cordatum, P. schiedeanum, P. chilense, P. falciforme and Blechnum varians are included. A biometric analysis of sporangia and spores, important taxonomic structures in the distinction of ferns, has been carried out. The data were subjected to a one-way analysis of variance and a discriminant analysis. In addition, the spores were observed under a scanning electron microscope to study their ornamentation. Of the characters we have studied thickness of the arcus, number of cells in the arcus, number of cells in the hypostome and major equatorial diameter of the spore have statistically supported taxonomic significance and are therefore useful for species separation. 
 


Introduction
Blechnaceae is a subcosmopolitan fern family, with about 265 species (PPG I 2016). The family consists of 25 genera (Gasper et al. 2016Molino et al. 2019). Among these, Parablechnum is the most diverse, comprising approximately 65 species with a pan-tropical distribution ( Fig. 1), predominantly in the southern hemisphere .
In the case of P. cordatum, P. schiedeanum, and P. chilense, there has been disagreement about whether they were co-specific (Tryon & Stolze 1993), or whether they were distinct species (De la Sota 1972;Moran 1995;Rodríguez 2015;Gasper et al. 2017), so several authors have indicated the need for a monographic study of the group Moran 1995;Tryon & Stolze 1993). Recent molecular phylogenetic studies suggest that they could correspond to three different species (Vicent 2017). Tryon & Stolze (1993) included P. chilense under the synonymy of P. cordatum, and this same criterion is adopted by Prada et al. (2008). There are authors who support the separation of both species based on morphological characteristics, such as Rodriguez (2015) who argues that P. cordatum and P. chilense differ in the shape and distribution of the scales of the petiole and rachis, the aerophores, and the sporal morphology.
As for P. falciforme, Moran (1995) maintains that it may be co-specific with P. schiedeanum. On the other hand, Blechnum varians appears as a synonym for Parablechnum schiedeanum according to Tropicos (2020), following the criteria of Mickel & Smith (2004), who believe that it is the same species. According to Moran (1995) Blechnum varians apparently differs from P. schiedeanum only by its dark axes and flattened or very sparse and irregular pinna margins, suggesting that it could be the same species since there is not so much difference.
To clarify the taxonomy of this group, a more extensive study of the genus is necessary, and among the characters that have been revealed to be of taxonomic importance, spores and sporangia are of special interest.
Spores have been widely used as a taxonomical character in ferns (Tryon & Lugardon 1991;Tryon & Tryon 1982). In particular, ornamentation, wall structure, size and shape of the spores are of high taxonomic value characters, since they are relatively constant for each species (Barrington et al. 1986(Barrington et al. , 2020Lugardon 1974;Tryon & Lugardon 1991).
Particularly, spores have come out as important characters within the family Blechnaceae, and so do the sporangia (Melo da Silva et al. 2019; Molino et al. 2020;Moran et al. 2018;Prada et al. 2016;Passarelli et al. 2010).
Thus, the aim of this work is to provide new data that help to elucidate the infra-generic taxonomy of the species complex Parablechnum cordatum through the biometry of spores and sporangia.

Biometric Analysis
A biometric analysis of the five taxa was carried out, using at least two individuals from each of them. The specimens analysed can be found in the Appendix.
For the optical microscopy observations, the samples were mounted directly in water and photographed with a Nikon Eclipse Ci microscope with a Nikon DS-Fi2 camera. From the photographs taken, the desired characters were measured with the Piximètre software.
At least 15 measurements of the following sporangial characters were taken on each individual: height and width of the capsule; number of cells and thickness of the arcus; number of hypo and epistomial cells; width of the cells of the upper and lower lip and length of the pedicel and rosette. All the terms used regarding the sporangia follows what is stated in Molino et al. (2020). For the length of the rosette, a minimum of 10 measurements were taken per individual.
Spores were studied by optical microscopy to establish their dimensions in a total of 11 individuals, mounted in water without previous treatment. In every individual, 15 measurements of the polar and equatorial major diameter were taken, excluding the perispore.
From the same individuals, sporal ornamentation was studied by scanning electron microscopy. The samples were mounted in a sample holder with carbon adhesive, metallized and observed in a scanning electron microscope (SEM) JSM 6400 JEOL operating at 20 kv. The observations were made at the National Center of Electronic Microscopy (CNME) of the Universidad Complutense de Madrid. Two variables were chosen in relation to ornamentation: small (< 10 µm) or large (>10 µm) areolae, and presence or absence of filamentous processes.

Statistical analysis
A biometric data matrix has been constructed using the measurements of individuals for each character. Using the SPSS software, the normality of the data was checked, doing a K-S test, followed by a Lilliefors' significance correction. Once the normal distribution of the variables was checked, in the following statistical analyses, parametric tools were used. In all cases, α=0.05 has been used as a criterion.
With the same software an ANOVA was made for each one of the mentioned characters with a Tukey's Post-Hoc HDS test to check if the difference between the means of the variables was significant.
Using the Statgraphics software, a discriminant function analysis was performed in 2 phases, first, using two dichotomous variables (small or large areolae, presence of filaments or not) that allowed us to differentiate 100% between 3 groups. In a second phase, the discriminant analysis uses the continuous morphological variables producing a discriminant function that allows separating the species of each group according to the most explanatory variables.
According to the ornamentation of the spores of each species, 3 groups could be distinguished: Group 1, formed by spores with small areolae, up to 10 μm and with filaments on the areolae surface, corresponding to P. chilense and P. falciforme; Group 2, which includes spores with large areolae, up to 30 μm and without filaments, corresponding to P. cordatum and P. schiedeanum, and Group 3, composed by spores with large areolae and with filaments in their central area, where only B. varians is found (Fig. 2).
These dichotomous variables have made it possible to differentiate one of the species in the complex, B. varians. In order to separate the taxa of groups 1 and 2, the analysis of the continuous variables was used.

Biometric analysis of characters
The biometric data for the characterization of the sample are shown in Table 1. With the discriminant analysis carried out in groups 1 and 2 resulting from the morphological analysis of the spores, the standardized coefficients of the discriminant function have been obtained for each group, which indicate the most important variables to separate P. chilense and P. falciforme in the first case, and P. cordatum and P. schiedeanum in the second. The results of the analysis are shown in Table 2, which highlights the most explanatory variables for the function in each group.
Of the five most significant variables to separate both species (Table 2), the three with higher absolute values were chosen for Group 1 to make the separation of P. chilense and P. falciforme easier. The width of the upper lip, the number of hypostomial cells and the thickness of the arcus showed a percentage of cases correctly separated of 100%, that is, with these three variables and their respective measures, there was no confusion between the mentioned taxa because these measures were sufficiently different so that none of them was confused, allowing the total separation of them. This is shown in the scatter plot in Fig. 3, where the separation of both species is seen with two of the most explanatory variables (width of the upper lip and thickness of the arcus).
For Group 2, of the five most explanatory variables (Table 2), the three with the highest absolute values were chosen to facilitate the distinction of P. cordatum and P. schiedeanum. The number of cells in the arcus (Fig. 4), the equatorial diameter of the spores, and the length of the rosette (Fig. 6) showed a percentage of well separated cases of 92.98% (4 of 57 measurements were not correctly separated). This can be seen in the scatter plot of Fig. 5, in which certain overlaps between individuals of both species are seen with the two most explanatory variables (equatorial spore diameter and number of cells in the sporangial arcus).
To homogenize the distinction of the species in this complex, the discriminant function has been calculated for both groups using the same variables, that is, the two most explanatory variables of Group 1 (the width of the upper lip and the number of cells in the hypostome) and Group 2 (the number of cells in the arcus and the equatorial diameter of the spores) have been chosen.
For Group 1 the percentage of well separated cases is 98.67% (1 out of 75 measurements have not been correctly classified), while for Group 2 the percentage is 90.74% (5 out of 54 measurements have not been correctly classified).

Discussion
Spores are a character of great importance when making taxonomic classifications in ferns (Barrington et al. 1986(Barrington et al. , 2020Tryon & Lugardon 1991;Tryon & Tryon 1982). Specifically, in the case of the family Blechnaceae several studies have been made concerning the spores (Melo da Silva et al. 2019; Moran et al. 2018;Pasarrelli et al. 2010), and recently others that allow us to discriminate species and genera within this family using, in addition to the spores, the information concerning the sporangia (Molino et al. 2020).
The differences in spore ornamentation, in terms of areolae and filaments, have been mentioned in both Parablechnum schiedeanum (Moran et al. 2018) and P. falciforme (Passarelli et al. 2010), coinciding our results with those presented in these publications. As for P. cordatum, its ornamentation was described by Prada et al. (2008) and Melo da Silva et al.
(2019), who indicated that the areas between ridges present filaments or that they present discrete filaments dispersed throughout the surface, respectively. Rodríguez (1970) described the spores of P. chilense, with perforations and evident laesura. Later this same author (Rodriguez 2015) highlighted that P. chilense and P. cordatum differ in the scales of the petiole, aerophores and palynology, but did not specify what the differences are. In our work we have been able to verify that P. cordatum does not present filaments on the surface of its spores and, therefore, the presence of filaments in P. chilense and its absence in P. cordatum is one of the most evident characteristics that allows differentiating both species.
It is not registered in any previous work the ornamentation of the spores of B. varians probably because it has been considered as a synonym of P. schiedeanum (Mickel & Smith 2004), however, according to our results B. varians presents big areolae and filaments in the surface of its spores, unlike the observed in those of P. schiedeanum.
In this work, it has been provided a statistical support related to the biometry of certain characters such as the width of the cells of the lip or the thickness of the arcus of the sporangia that has not been documented before. Likewise, it has been possible to separate the five taxa inside the complex, through the statistical comparisons of the means among the species with a discrimination of more than 90% in all the cases.
Despite having made a homogenization to facilitate the separation of taxa belonging to this complex, it is important to highlight the eight variables that most explain the separation between groups and within groups, because they provide much information about each taxon. Even so, for a simple identi-fication, using only the four variables chosen we can do it in a correct way in more than 90% of the occasions.
Respecting the species that we treat in this work, only the sporangia of Parablechnum cordatum have been described before (Prada et al. 2016), where it is indicated that they are sporangia with (16) 17 (20) cells in the arcus, 4-6 in the epistomium, 3-5 in the hypostomium and 2-4 in the stomium (lips), wide pedicel that narrows abruptly in the rosette, whose cells are narrower, 1600 μm long and 160 μm wide. In our work, we have been able to verify that clear differences are observed in the number of cells in the epistomium and in the length of the pedicel, which are 3-4 and 264-488 μm respectively. The mentioned differences perhaps are due to that these authors treated P. chilense like synonym of P. cordatum, reason why the values that offer are nearer to which here we have obtained for P. chilense.
The sporal size values of P. chilense presented by Rodríguez (1970) are higher than those obtained in our work, and from them it is deduced that spores would tend to have a contour that would tend more to be circular than ellipsoidal in polar view; probably these differences are explained because in the measurements provided by that author the perisporium is included. The data of the spores and sporangia make clear that P. chilense presents differences with respect to P. cordatum, which agrees with the molecular data of Vicent (2017) and the morphological observations of Rodríguez (2015).
The rest of the biometric characters that have been studied in this work have shown to have an interesting diagnostic value, for which they should be considered in future studies, which was already shown by Prada et al. (2016) in relation to the number of cells in the arcus, among other characters.
It differs from Parablechnum schiedeanum in its atropurpureous petiole and rachis, crisped pinna margins and spores with filamentous micrornamentation.
With the results obtained, we elaborated a dichotomous key using the most significant characters of the statistical analysis that allows to distinguish in a simple way the species of this complex.
Key to the species within P. cordatum complex 1. Perisporium without filaments on the areolae surface 2 Perisporium with filaments on the areolae surface 3 2. Rosette length and equatorial spore diameter < 60 µm P. schiedeanum Rosette length and equatorial spore diameter >60 µm P. cordatum 3. Perisporium with large areolae, up to 30µm P. varians Perisporium with small areolae, up to 10 µm 4 4. Rosette length and ring thickness < 60 µm P. chilense Rosette length and ring thickness > 60 µm P. falcifome Tables   Table 1 Table 2. Standardized coefficients of the discriminant function for both groups. The green shaded boxes are the most explanatory variables for the function. Tabla 2. Coeficientes estandarizados de la función discriminante para ambos grupos. Las casillas sombreadas en verde son las variables más explicativas para la función.