Periphytic Actinella Lewis ( Ochrophyta , Bacillariophyceae ) species from an Environmental Protection Area in the Brazilian Amazon Espécies perifíticas de Actinella Lewis ( Ochrophyta , Bacillariophyceae ) de uma Área de Proteção Ambiental na Amazônia Brasileira

Aim: The aim of this study is to present the Actinella species found in periphytic samples collected from an Environmental Protection Area in Santarém (PA, Brazil), to comment on their morphology by comparing them with existing records, and to increase knowledge of the distribution of the genus in Pará State. Methods: Ten periphytic samples were collected from seven sampling stations in the Alter do Chão Environmental Protection Area (Santarém, Pará, Brazil) in October 2015 and February and July 2016. Dissolved oxygen, pH, conductivity and water temperature were measured in the field at sites from which samples were collected in 2016. Samples were oxidized and analyzed by light microscopy. Results: The sampled water bodies presented low conductivity and an acidic pH. We identified 12 Actinella taxa, several of which had originally been described from samples collected from the Amazon in the mid-20th century, mainly in the lower Tapajós River region. Actinella rionegrensis is recorded for the first time in the State of Pará, outside of its type locality (Negro River, near Manaus, AM, Brazil), and the species habitus (wisp-shaped colonies) is recorded for the first time. Conclusion: Our results indicate that the Environmental Protection Area has been effective in conserving the aquatic ecosystem, since Actinella species identified in the mid-20th century are still present. Taxonomic issues for some species of this genus need clarification and revision, since overlapping diagnostic features occur among species and with species belonging to the genus Eunotia.


Introduction
The order Eunotiales Silva belongs to the subclass Bacillariophycidae D.G. Mann, Class Bacillariophyceae Haeckel, sensu Ruggiero et al. (2015).They are the only group of diatoms whose valves have both a raphe system and rimoportulae.There is no central nodule, the raphe is placed distally and reduced, and in most genera it is mainly located in the mantle (Round et al., 1990;Siver et al., 2015).To date, 12 genera have been proposed (more recently Perinotia Metzeltin & Lange-Bertalot, Bicudoa Wetzel et al., Eunotioforma Kociolek & Burliga, and Actinellopsis Taylor et al.), some of which are not widely recognized (Siver et al., 2015).
Periphytic Actinella Lewis (Ochrophyta, Bacillariophyceae) species… 19 of which were recorded in Pará State, mostly collected from the lower Tapajós River area.Metzeltin & Lange-Bertalot (1998, 2007) described 377 new diatom taxa from tropical environments, mostly from samples collected in Colombia, Venezuela, Guyana, and Brazil.In those studies, the authors drew attention to the diversity of the Tapajós River aquatic system, particularly Jurucuí Lake that is the type locality for several new species of Actinella proposed by Metzeltin & Lange-Bertalot (1998, 2007) and Kociolek et al. (2001).They also pointed out the considerable species richness of the Order Eunotiales, totaling 194 and 24 species of the genera Eunotia and Actinella, respectively (16 of these latter were proposed as new species).
Almost 70 years have passed since the collections of Rudolf Braun and Harald Sioli (Braun, 1952;Sioli & Soares, 2006) from the Santarém region (PA, Brazil), samples from which were used to describe most of the known South American Actinella species.We collected periphytic samples from the lower Tapajós River region as part of an investigation of the periphytic diatoms in the Environmental Protection Area of Alter-do-Chão (APA Alter do Chão, Santarém, PA, Brazil).We aimed to extend our knowledge of tropical South American Actinella species by updating and expanding records for the State of Pará.Here, we describe and illustrate the taxa, we comment on their morphologies and habitus, and we address some taxonomic issues.

Material and Methods
Study Area -The Environmental Protection Area of Alter do Chão (APA Alter do Chão) was created in 2003.It covers an area of 16,180 ha and is located approximately 35 km from the city of Santarém, in Pará State, Brazil.The protected area includes several streams ('igarapés') and floodplain lakes connected to the Tapajós River (a 'clearwater river' according to Sioli's Amazonian river typology, Sioli, 1984), including the lakes Verde and Jurucuí.Verde Lake has a surface area of approximately 445 ha and is situated at the right bank of the Tapajós River (Figure 1).
Sampling and analysis -The 10 periphytic samples used in this study were collected from seven sampling stations in October 2015 and February and July 2016 (Table 1).We collected leaves, petioles, stems or whole submerged macrophytes (in the case of small plants), submerged tree bark, and superficial sediments from the littoral zone of Verde Lake and the banks and beds of its associated streams.To collect surface sediments, we gently pushed an inverted Petri-dish against the submerged surface and slid a rigid plastic square underneath it.Samples were stored in bottles or Ziploc  plastic bags.In the laboratory, samples were conditioned in flasks and fixed with 4% formaldehyde solution.
During the February and July 2016 sampling campaigns, we measured in situ pH (pH system model 60, YSI, Ohio, USA), water temperature and dissolved oxygen (oximeter model 55, YSI, Ohio, USA), and electrical conductivity (Ecosense model EC 300A conductivity/temperature, YSI, Ohio, USA).To equalize sample sizes for oxidation, portions of macrophytes were cut into small pieces and placed in a graduated test tube up to the mark of 3 ml (3 cm 3 ) (modified from Lim et al., 2001).The same volume was used for sediment samples.Samples of the periphyton were extracted by brush scraping (for macrophytes) or by shaking in distilled water (both macrophytes and sediment).Aliquots of 2 ml were oxidized according to Von Stosch (1970).Permanent slides were made in duplicate using Naphrax  (R.I. = 1.74) as a mounting medium.
The slides were observed under a Zeiss Axio Scope A1 light microscope (LM) with a 100x oil immersion objective at a magnification of 1000x.Specimens were photographed and measured using the ZEN 2012 (blue edition) software and image capture system.The widths were measured at the midsection of the valves.We adopted the classification for higher taxonomic levels (from phylum to order) proposed by Ruggiero et al. (2015), and identifications to specific and infra-specific level were based on several studies of flora and both classic and recent publications (van Heurck, 1881;Hustedt, 1952Hustedt, , 1965;;Metzeltin & Lange-Bertalot, 1998, 2007;Kociolek et al., 2001;Ferrari et al., 2007;Melo et al., 2010).Samples are currently housed in the HSTM-Algas herbarium at the Federal University of Western Pará (UFOPA).

Results and Discussion
The low values of electrical conductivity and the acidic pH we recorded from our water samples indicate oligotrophic water conditions in the lower Tapajós region (Table 1), which are suitable for Actinella and for the Order Eunotiales in general (Round et al., 1990;Montoya-Moreno & Aguirre-Ramirez, 2013).

SS
Studied material (n=15): HSTM-Algas 9490, 9522, 9521, 9522, 9535, 9537, 9538, 9552.Comments: Measurements of our specimens are concordant with those presented in the protologue.In the identification key provided by Kociolek et al. (2001), the distinctive feature of this taxon is its concave headpole margin, although this was not mentioned in the original description of the species.We consider this feature, as well as the taxon's strong dorsal convexity, as differentiating it from other similar taxa, such as A. crawfordii and A. lima.
Comments: The striae density is not given in the original description.Metzeltin & Lange-Bertalot ( 2007) noted a probable mistake in the scale presented in the protologue illustrations (Kociolek et al., 2001).Considering the scale given with Fig. 113, the specimen is approximately 2.3 μm wide and presents around 25 striae in 10 μm, whereas specimens shown in Figs.114 and 116 could be estimated as having 16 and 18 striae in 10 μm, respectively.These latter numbers are consistent with the specimens identified in our study, as well as with those recorded by Melo et al. (2010), supporting that there is indeed a mistake with respect to the scale bar provided with the specimen shown in Fig. 113.Therefore, we consider the striae density for A. gracilis to be in the range of 16-18 in 10 μm.
Comments: Measurements, striae density and shape of our specimens are concordant with those presented in the protologue (Van Heurck, 1881) and in Metzeltin & Lange Bertalot (1998), Kociolek et al. (2001) and Periphytic Actinella Lewis (Ochrophyta, Bacillariophyceae) species… Melo et al. (2010), all of them from Amazonian samples.Bicca et al. (2011, Figs. 2 and 3) and Dunk et al. (2016, Figs. 2d -f ) present specimens of A. guianensis with non-apiculate, less dilated headpoles, from samples collected outside the Amazon (Southern and Central Brazil, respectively).Additionally, the headpole and the footpole of those specimens are almost of the same width, unlike those from Amazonian samples.Further investigation is needed to determine whether they are morphotypes or distinct species.
Comments: The analyzed specimens have a more pronounced apical projection than those illustrated by Kociolek et al. (2001), in which this structure hardly projects beyond the valve body and is more rounded.However, the morphology of our specimens agrees with the specimens illustrated by Metzeltin & Lange-Bertalot (2007) (Figure 28: 3-8).This taxon is quite similar to A. gracilis, mainly in relation to the valve outline and the headpole shape, but it differs from it by the valvar width being ≥ 4 μm (Kociolek et al., 2001), which is the principle feature we used to distinguish these two taxa.
Comments: Actinella pseudohantzschia is related to A. eunotioides Hustedt, but differs from it by the pattern of the striae, its slightly larger width, and the presence of conspicuous spines (even under LM) at the valvar margins.In the protologue, Hustedt (1952: 135) observed the presence of spines (7-10 in 10 μm) in A. eunotioides.However, when comparing these two species, Metzeltin & Lange-Bertalot (2007) stated that spines were absent from the samples collected in Guyana and Brazil (including the type material), and provided details of length, width and striae density that do not match those provided in the protologue for A. eunotioides.In an amendment of Hustedt's original description, Kociolek et al. (2001) report the presence of spines in the margins, but in their commentary of observations by scanning electron microscopy, they say that they are absent.
Eunotia roland-schmidtii Metzeltin and Lange-Bertalot is also similar to Actinella pseudohantzschia, but is longer, and its width and striae number overlap with those of A. eunotioides.Individuals with or without conspicuous spines have been illustrated (Metzeltin & Lange-Bertalot, 2007: 457).The authors of that report expressed uncertainty regarding whether E. roland-schmidtii belonged to the genus Eunotia or to Actinella, since longer specimens of Actinella tend to be isopolar and several Actinella species lack spines.Actinella disjuncta Metzeltin and Lange-Bertalot is another species similar to A. pseudohantzschia, but it generally is shorter and exhibits greater striae density.
Apart from the individuals we illustrate here (whose characteristics match the description of A. pseudohantzschia), we found specimens from several samples that simultaneously present morphometric characteristics of A. pseudohantzschia, A. eunotioides and E. roland-schmidtii.In general we found that the larger the specimen the harder it was to assign species identity to it due to progressively subtle heteropolarity and overlapping morphometric features.We propose that more detailed ultrastructural analyses of populations are needed to establish distinct species boundaries, to differentiate lifecycle stages, and to reveal intraspecific morphological variation.
Comments: Morphometric characteristics of the specimens we examined are in agreement with those presented for the species (Hustedt, 1952;Kociolek et al., 2001).Metzeltin & Lange-Bertalot (2007) reported that they observed large morphological heterogeneity in the samples of Hustedt's collection from Brazil, stating that the morphological groups they delineated may constitute distinct species or subspecies.In that same work, the authors proposed two new species related to A. peronioides, i.e. A. superperonioides Metzeltin & Lange-Bertalot (from the same sampling location as the A. peronioides type material) and A. subperonioides Metzeltin & Lange-Bertalot (representing samples from Venezuela, i.e. not the type locality of A. peronioides).Compared to A. peronioides, A. superperonioides presents a larger valve length (about 100 μm compared to <75 μm for A. peronioides), a larger footpole (3.5-5.0 μm compared to <2.3 μm for A. peronioides), a larger mid-valve width (4.0-5.0 μm, compared to <3.5 μm for A. peronioides), more arched valves, subcapitate apices and lower striae density, as well as less capitate headpoles.A. subperonioides (see Figures 4f-g) differs from A. peronioides by having narrower rather than capitate headpoles.Some related morphological groups remained undetermined (Figure 46: 18-32 in Metzeltin & Lange-Bertalot, 2007).
Observations of non-oxidized samples revealed pairs of individuals adhered to the substratum through a branched (bifurcated) mucilaginous stalk (Figure 5b), as reported for A. aoteoaroaia Lowe et al., illustrated in Sabbe et al. (2001).This latter taxon also exhibits a very similar girdle outline compared to A. peronioides.
Comments: As they did for the species group related to A. eunotioides, Metzeltin & Lange-Bertalot (2007) proposed the separation of A. peronioides into at least two more species (A.subperonioides and A. superperonioides, with some indeterminate morphotypes).Some individuals that we analyzed possess a valve contour that is very similar to the undetermined specimens shown in Metzeltin & Lange-Bertalot (2007, Figure 46: 24-32), which are between 12.3-41.5μm in length, 3.8-4.6μm (headpole) and 3.1-3.8μm (footpole) in width, and have 18-19 striae in 10 μm.Relative to the new species proposed by Metzeltin & Lange-Bertalot (2007), our specimens more resemble A. subperonioides (Metzeltin & Lange-Bertalot, 2007, Figure 46: 1-10), but have a greater striae density, Periphytic Actinella Lewis (Ochrophyta, Bacillariophyceae) species… which is why we designate them A. aff.subperonioides.With respect to the A. peronioides species complex, again there are problems concerning overlapping measurements and the existence of a morphological continuum among specimens.This makes it difficult to assign specimens to species, as discussed above for the A. eunotioides species complex.Thus, it remains unclear if A. peronioides should be differentiated into more than one species.
Comments: This species presents a unique undulating pattern on both margins, so it can hardly be confused with any other currently described taxon.Metzeltin & Lange-Bertalot (1998: 242) proposed a 'natural constant' concerning the number of undulations on the margins of A. mirabilis, i.e. that the number of undulations on the dorsal side is always two more than that on the ventral side.Our analysis supports this claim, since the specimen shown in Figure 6a exhibits nine undulations on the dorsal margin and seven on the ventral margin.Observations of non-oxidized epiphytic samples revealed solitary individuals adhered to the substratum by a thick mucilaginous stalk (Figure 5a), confirming the sessile habitus of this taxon, as noted by Uherkovich (1986) in Kociolek et al. (2001).
Comments: Whether this isopolar taxon should be included in the genus Actinella is another example of the difficulty of establishing and applying morphological criteria that can clearly separate the genera Eunotia and Actinella.Observations of non-oxidized samples revealed several frustules joined at their footpoles (forming rosette colonies, sensu Rimet & Bouchez, 2012) that were adhered to the substratum (Figure 5c).Adherence to the substrate at one of the poles is common among species of Actinella, whereas this is rare in the genus Eunotia.We present the first record for this taxon outside its type locality and for Pará State, as well as the first description of its colony morphology.
Comments: Valve length reaches 400.0 μm in the protologue for A. robusta.We did not find records of such long specimens in other studies; maximum valve length is about 184.0 μm in Kociolek et al. (2001, p.139, Figs 36-39), 216.0 μm in Melo et al. (2010), and 173.6 μm in this study.

Conclusion
Our investigation of Actinella species from the lower Tapajós River region revealed that this flora has not changed very much since collections made in the mid-20 th century, indicating that the Environmental Protection Area is effective in terms of conserving the aquatic ecosystem.The only species we found that had not previously been reported for this region was Actinella rionegrensis, which was initially described from samples collected in the Negro River, a 'blackwater' river (according to Sioli, 1984) and one of the most important tributaries of the Amazon River.However, if the taxonomic distinctiveness of A. subperonioides is confirmed, it will be the first record of this taxon in Brazil (and outside its type locality of Venezuela).
We found 12 of the 25 taxa presented in the study carried out by Melo et al. (2010), who summarized the occurrence of Actinella in the Brazilian Amazon.It is notable that from only 10 samples analyzed in this study, we encountered about 50% of the Actinella diversity reported in previous studies for this region, including those by Metzeltin & Lange-Bertalot (1998, 2007), wherein about 500 slides from tropical South American samples were investigated.
We did not find the 13 species that have previously been reported for the Brazilian Amazon.Of these, we believe that neither A. punctata (that is the typus generis) nor A. tasmaniensis occur in the Brazilian Amazon, even though both of these taxa were reported in the study by Uherkovich & Franken (1980), but the authors did not provide any illustrations.That study represents the only report for A. tasmaniensis outside Australasia (Williams & Kociolek, 2016).We believe that the specimens identified as A. tasmaniensis and as A. punctata by Uherkovich & Franken (1980) were probably confused with other similar species, as is likely the case for reports of A. punctata in other countries (Kociolek et al., 1997).Actinella punctata is commonly found in the northern regions of the Holarctic and has only been reported in South America from a lake in Colombia c. 3000 m above sea level (Metzeltin & Lange-Bertalot, 2007).Alencar et al. (2001, Fig. 33) also reported A. punctata from the Brazilian Amazon, but their illustrated specimen is actually A. siolii.The remaining species not registered in this study but that have been previously reported for the Brazilian Amazon are also rarely recorded in other studies or have only been found once at their type localities.
The gradient of morphological characters found among some taxa, together with overlapping character measurements, makes assigning taxonomy for some specimens difficult, as has been found for the A. brasiliensis species complex (comprising A. brasiliensis, A. hustedtii Kociolek, A. spauldingae Metzeltin & Lange-Bertalot, A. curvatula, A. lima, A. crawfordii and A. gracile) and the A. peronioides species complex (comprising A. peronioides, A. subperonioides and A. superperonioides).It is even difficult to assign certain taxa to either the Actinella or Eunotia genera, as is the case for A. pseudohantzschia, A. lange-bertalotti Kociolek, A. eunotioides and E. roland-schmidt.Thus, a more detailed revision of this group is necessary, using morphological, ultrastructural, molecular, and biogeographical data.Observations of species forms throughout their lifecycle and their habitus in fresh samples may help.Actinella commonly attaches itself to the substrate through mucilage pads or stalks that grow from the footpole.This genus is either solitary or forms rosette colonies, which contrasts with the long ribbon-like chains observed in Eunotia in which frustules are bound at their valve surfaces or at their poles to form filamentous branched colonies (personal observation).

Figure 1 .
Figure 1.Maps of the study area showing the locations of the sampling stations (a), in the lower Tapajós River region (b) of Pará State (c).
Valves heteropolar, slightly arcuate, margins parallel.Headpole broadly rounded, with median apical projection, which is progressively displaced to the dorsal side as the valvar length increases; footpole narrowly rounded.Length