Morphological study of the sclerites of the species Renilla muelleri Kölliker, 1872 (Anthozoa, Octocorallia, Pennatulacea)

. Some octocorals (Cnidaria, Anthozoa) can produce a sub-millimeter to a millimeter-range skeleton of calcium carbonate, known as sclerites. Sclerites have diverse morphological characters, such as size, and shape features that give them relevance as taxonomic characters at different levels. Renilla Lamarck, 1816, is a genus of the order Pennatulacea (Octocorallia), in which colonies are characterized by two distinct zones: a peduncle, which anchors them to the substrate, and a rachis that supports the polyps. In the case of Renilla muelleri Kölliker, 1872, prior research established the presence of similar sclerites in both the peduncle and the rachis. Nevertheless, the potential morphological variations of these sclerites among colonies and within different colony regions have yet to be assessed to determine the stability of these characteristics at the species level. This study aimed to describe and compare the external microscopic morphology and biometry of the sclerites of Renilla muelleri, enhancing their identification and assessing their consistency within the species. Sclerite composition was examined across the entire colony, and we analyzed length variation among colonies using generalized linear models (GLM) and within colony zones using generalized linear mixed models (GLMM). Additionally, the external microscopic morphology of all sclerites was examined through scanning electron microscopy (SEM). Based on size, two types of sclerites, namely large and small, were identified within the colonies. Both types exhibited significant size differences among colonies. Furthermore, the large sclerites displayed notable variations between zones, with those in the rachis being the largest and holding the highest rank within the colonies. In terms of external microscopic morphology, the sclerites exhibited considerable variability, making it challenging to establish clear groupings based on these characteristics. Based on the obtained results, it can be concluded that both the biometry and external microscopic morphology of sclerites do not exhibit consistency as characteristics in Renilla muelleri highlighting the ambiguity in defining Renilla species.


INTRODUCTION
The size, and external microscopic morphology of sclerites are pivotal features in the taxonomy of organisms within the subclass Octocorallia (Cnidaria: Anthozoa), as they serve as diagnostic characteristics (Sethmann et al., 2007).Nonetheless, the identification and classification of different sclerite types pose a challenging task due to the considerable variation in size and shape observed both within and between species (Lewis & von Wallis, 1991;Vargas et al., 2010;Mastrototaro et al., 2015;Gutiérrez-Rodríguez et al., 2009;West et al., 1993;Everton et al., 2015).Some au-thors showed that SEM analyses can provide information on the morphological differentiation between them and can be considered diagnostic at the species level (Sánchez, 2007).Aharonovich & Benayahu (2011), Benayahu et al. (2018), and Halász et al. (2019) attributed this morphological difference to the modes of growth and aggregation of calcite crystals.
The history of the genus Renilla Lamarck, 1816 (Pennatulacea) has been marked by challenges in species description.It previously included 13 species, of which only 6 are currently considered valid.Incomplete descriptions and the difficulty in identifying reliable taxonomic characteristics have contributed to these taxonomic issues, exemplified by cases such as Renilla chilensis Philippi, 1892 (considered a nomen dubium) (Pérez & Zamponi, 1999).Prior literature has suggested that the sclerites of the genus Renilla consist of three-flanged rods and exhibit similar lengths in both the peduncle and rachis (Pérez & Ocampo, 2001).Renilla musaica Zamponi & Pérez, 1995 represent the unique exception showing variations in the lengths of sclerites between rachis and peduncle.Renilla muelleri Kölliker, 1872, is among the most well-documented species within the genus and is widely distributed along the Argentine coast (Zamponi & Pérez, 1995).Previous research, as conducted by Clavico et al. (2007), has specifically examined variations in sclerite sizes in R. muelleri while also established links between these variations and the depth at which samples were collected.While sclerites have been recognized as diagnostic features between species, it is essential to consider the possibility of morphological variation within a given species.Based on quantitative data, the significance of these features and their potential utility for future taxonomic purposes remains to be comprehensively evaluated.This study aimed to describe and compare the microscopic external morphology and biometry of Renilla muelleri sclerites in order to facilitate the identification and comparison of these structures within the species.Additionally, the study aimed to assess the consistency of sclerites as a taxonomic character at the species level.

Material studied
The studied specimens were identified as Renilla muelleri following the morphological characteristics presented by Zamponi & Pérez (1995).Colonies have a horseshoe or kidney-shaped form.They exhibit a relatively short peduncle.Autozooids feature five calycinal teeth.In the dorsal section, the rachis is completely covered with zooids, while the ventral part of the rachis and peduncle display a deep violet coloration (Fig. 1).According to the original description (Zamponi & Pérez, 1995), R. muelleri presents peduncle sclerites of the same length as those of the rachis.This characteristic will be discussed below, as the study of sclerite lengths is the focus of this investigation.On the other hand, the original description of R. muelleri identifies the cnidocysts: microbasic amastigophores, holotrichous isorhizas, atrichous anisorhizas, and macrobasic mastigophores (Zamponi & Pérez, 1995).However, in the specimens analyzed here, only one type of cnidocyst could be identified, probably atrichous isorhiza, although without certainty.

Protocol of sclerite preparation
To obtain sclerite samples, 5 mm³ sections of tissue were carefully extracted with a scalpel from both the rachis and peduncle.Subsequently, these samples were subjected to treatment with sodium hypochlorite to facilitate sclerite dissociation.Finally, ethanol and distilled water were used to clean the sample.

Sclerite composition and biometry
As the data originated from different collection sites, we kept the separation by distinguishing between the "Undine" (referred to as "U") and "Holmberg" (referred to as "H") campaigns for our analysis.We identified the distinct types of sclerites according to their size, and their location (peduncle and rachis), and these were subsequently measured and photographed under an optical microscope at 40X magnification.To aid in these tasks, we employed the Leica Application Suite EZ software (version 3.4.0,Leica Microsystems,Wetzlar,Germany).
The normality of sclerite length was evaluated by means of a Shapiro-Wilk test (α = 0.05), applied to the residuals obtained from a linear model assumed to have a normal distribution.If normality was not supported, a Generalized Linear Model (GLM) with a gamma distribution for error terms was employed as an alternative approach.The adequacy of these models was assessed through graphical methods, including Quantile-Quantile (Q-Q) plots, depicting the relationship between the residuals and the theoretical quantiles of the model, as well as scatter plots illustrating the deviation of residuals about the model's predicted values.

Analysis of sclerite type length variation among colonies
As our data originated from two collection sites, we conducted an assessment of their influence on sclerite length variations among colonies.To account for potential dependencies, a Generalized Linear Mixed Model (GLMM) was fitted, with the variable 'Campaign' included as a random effect.The model was formulated as follows: To evaluate the statistical significance of the random effect associated with the variable 'Campaign, ' we performed a comparison between a Generalized Linear Model (GLM) and the GLMM using ANOVA (α = 0.05).The Akaike Information Criterion (AIC) was employed to determine the model that best fit the data.Subsequently, differences between the β₁ coefficients of the colonies were assessed using a t-test (α = 0.05) to ascertain disparities in sclerite length among the various colonies for each sclerite type, separately for both peduncle and rachis.

Analysis of sclerite length variation between peduncle and rachis
Given that the dataset comprised measurements from 10 colonies across two campaigns conducted at various depths, similar to the previous analysis, the variables "Campaign" and "Colony" were treated as random effects during the significance assessment in the analysis.The variable 'Zone' was included as fixed effect, assesing rachis and peduncle data.Consequently, the formulated model was as follows: To select the optimal model, the significance of each random variable was assessed by iteratively removing them one at a time, employing Akaike's Information Criterion.Subsequently, the best model was compared against a null model using ANOVA (α = 0.05).From the Generalized Linear Mixed Model (GLMM), the variance explained by the variables 'Colony' and 'Campaign' was examined.Subsequently, a t-test (α = 0.05) was performed on the coefficients β i of the GLMM to compare sclerite-type lengths between the peduncle and rachis zones.

Size of the colony
The estimation of colony size was conducted by measuring the area of the dorsal face of the rachis on scaled photographs through the utilization of the ImageJ software.To assess the potential correlation between colony size and sclerite size, a Spearman correlation test was employed, with a significance level set at α = 0.05.

External microscopic morphology analysis of sclerites
The sclerites derived from both the rachis and peduncle were prepared following the protocol outlined in Section "Protocol of Sclerite Preparations".Following preparation, the samples were affixed to a scanning electron microscope (JEOL JSM-646LV) slide surface using bifacial tape and subsequently coated with a gold-palladium layer for observation.Photomicrographs of the sclerites from the rachis and peduncle were captured to analyze the ultrastructural features of their external morphology.

Composition and Biometry of Sclerites
The composition and biometry of the sclerites exhibited similarity across both campaigns.Two discernible types of sclerites were identified in both the peduncle and rachis based on their size, labeled as "large" and "small." The size distinction between the two sclerite types was defined by a length threshold of 200 ± 3 µm according to the distribution of sizes in a kernel density plot (Fig. 3-5, Table 1).A total of 2,400 sclerites were measured (60 per type and zone of the colony).

Analysis of sclerite length variations among colonies
The variable "Campaign" was determined to be non-significant in the analysis and was subsequently excluded.The GLM model was selected for all cases (Table 2).According to the t-test for β i of the GLM, significant differences in sclerite length between colonies were observed for both types of sclerites in both the rachis and peduncle (Tables 3-6).

Analysis of sclerite length variations between peduncle and rachis
Normality tests for sclerite length data led to rejection in all datasets [rachis: small (p < 0.001), large (p = 0.004905), peduncle: small (p < 0.001), large (p < 0.001)].Consequently, a generalized linear model was fitted, demonstrating a good fit (Appendix: Figs.1-4).For small sclerites, the optimal model incorporated the variable "Colony" as a random effect.In the case of large sclerites, the model included both "Colony" and "Campaign" as random effects (Table 7).This suggests that the "Colony" variable had a significant effect on both large and small sclerites, while the "Campaign" variable was only significant for large sclerites.
The small sclerites exhibited no significant differences (p = 0.845) in length between those from the peduncle and rachis.Conversely, the large sclerites displayed significant differences in length (p < 0.001), and based on the estimated model values, those from the rachis (estimated = 338.53µm) were longer than those from the peduncle (estimated = 298.18µm).The variables "Colony" (standard deviation = 64.81)and "Campaign" (standard deviation = 73.55)accounted for a substantial portion of the model's variance, while the variance explained by the     residuals was low (standard deviation = 0.04) (Table 8).
The confidence intervals for the rachis sclerites were broader compared to those of the peduncle sclerites.

Correlation between sclerite size and colony size
The colonies collected in the "Undine" campaign were smaller in size than the colonies from the "Holmberg" campaign (Table 9).In the rachis, the large sclerite length showed a barely negative correlation (p < 0.001; rho = −0.28)with colony size; meanwhile, no significance was found for Table 3. Renilla muelleri.Rachis, sclerite small.Estimated values and confidence intervals derived from GLM, along with the p-value of the t-test for β₁ of the model.* Indicates significant differences for α = 0.05.The p-value for individual 1H is not displayed, as it was employed for comparisons."H" denotes colonies from the Holmberg campaign, and "U" corresponds to colonies from the Undine campaign.the small ones (p = 0.09).In the peduncle, the lengths of large sclerites had a negative correlation (p = < 0.001; rho = −0.42),and for small sclerites, the correlation was slightly positive (p < 0.001; rho = 0.20) (Figs. 6 and 7).

External microscopic morphology
In terms of the microscopic external morphology of sclerites, both the peduncle and rachis exhibited spindles.SEM observations revealed considerable variability in the external morphology of each sclerite type.Within the specimens, sclerites displayed predominantly smooth and regular margins, while others exhibited mostly irregular margins.All sclerites featured grooves along the longitudinal axis, some linear and/or spiral, complete from one end of the axis to the other and/ or incomplete, present only in a variable portion of the sclerite.These morphological characteristics were consistent across both sizes and in both areas of the colony in both campaigns.The external microscopic morphology variability encountered precluded discrimination or grouping based on these characteristics (Fig. 8).

DISCUSSION AND CONCLUSION
The analysis of sclerite composition in the octocoral Renilla muelleri, considering external microscopic mor-Figure 6. Stratified boxplot for large sclerites of Renilla muelleri by colony zone.Colonies are arranged on the x-axis in increasing order of size."H" denotes colonies from the Holmberg campaign, and "U" corresponds to colonies from the Undine campaign Figure 7. Stratified boxplot for small sclerites of Renilla muelleri by colony zone.. Colonies are arranged on the x-axis in increasing order of size."H" denotes colonies from the Holmberg campaign, and "U" corresponds to colonies from the Undine campaign Table 9. Size of Renilla muelleri colonies estimated from the rachis area (cm²) of the colonies."H" denotes colonies from the Holmberg campaign, and "U" corresponds to colonies from the Undine campaign.
The results highlight intraspecific variation in the length of both types of sclerite among colonies, a pattern consistent with findings from other studies on alcyonarian colonies (West et al., 1993;Gutiérrez-Rodríguez et al., 2009;Everton et al., 2015).Clavico et al. (2007) similarly observed size variation among colonies of Renilla muelleri, attributing it to phenotypic variability associated with colony depth.These observations reflect uncertainty regarding the size of sclerites in defining Renilla species.The descriptions of R. muelleri and R. musaica are very similar, and the terminology used to describe the sclerites is somewhat imprecise.For instance, phrases like "sclerites short and wide in the rachis" and "long and thin in the peduncle" used in the description of the species (Zamponi & Pérez, 1995), lack clarification on what constitutes short or long.Indeed, just two differences stand out between the descriptions of R. muelleri and R. musaica.Firstly, there is the issue of cnidocyst identification.We could not definitively identify any of the cnidocyst types mentioned in the descriptions of either species.Only one structure resembling a cnidocyst could be probably identified as an atrich isorhiza.This type of cnidocyst is mentioned in the original description of R. muelleri.Secondly, R. musaica is defined by exhibiting differences in sclerite length between the rachis and peduncle; while R. muelleri shows no differences.Our work reveals differences between the "large" sclerites from the rachis and peduncle in R. muelleri, with those from the rachis being longer.Zamponi & Pérez (1995) reported only one type of sclerite based on size in both species without distinguishing between "small" and "large".Figure 9 compiles and shows graphically the size of sclerites from R. muelleri obtained in this work and those from R. muelleri and R. musaica extracted from Zamponi & Pérez (1995).The values indicated by these authors for the sizes of the sclerites of both R. muelleri and R. musaica overlapped with the sizes found in this study for the "large" sclerites (Fig. 9).Additionally, it is observed that the "small" sclerites correspond to those found in the peduncle of a colony of R. musaica from Zamponi & Pérez (1995) (Fig. 9).These observations, added to the variability observed in sclerites of Renilla species (Clavico et al., 2007), raise doubts regarding the validity of R. musaica.It could potentially be considered a junior synonym of R. muelleri, although a detailed analysis of type material is required to clarify this matter.
The colonies from the "Holmberg" campaign were collected at a shallower depth than the colonies from the smaller "Undine" campaign and were also larger.The findings indicated that, for both the rachis and the peduncle, the reduction in colony size is associated with a slight increase in the size of the large sclerites, while such a trend is not observed for the small ones.
On the other hand, it was observed that the large sclerites exhibit variations in their sizes within colonies and between campaigns, with those from the rachis being larger than those from the peduncle.Clavico et al. (2007)   mary objective of this study was not to assess the potential cause of the variation, our results align with those of Clavico et al. (2007) only in the context of the large sclerites, where smaller colonies collected at greater depths displayed a slight increase in size.As a hypothesis, the larger size sclerites could potentially provide structural reinforcement to the colonies, serving either to withstand environmental pressure or to compensate for the fragility associated with smaller sizes, or perhaps both.However, this needs further investigation for validation.Variations in sclerite size have also been related to gradients of variation in predation pressure and the mechanical force of waves and currents that vary with depth (West, 1997).
Renilla muelleri exhibits simple sclerites visible with the naked eye under a light microscope, yet significant external morphological variability becomes apparent under SEM.Regarding ultrastructural characteristics, the observation of diverse external morphologies was possible, displaying a wide range of variation for each zone of the colony.This variability hindered discrimination and grouping based on these characteristics under optical microscopy.Consequently, the ultrastructure of external morphology, according to our results, does not appear to be a consistent character at a specific taxonomic level.However, it is worth noting that other authors, such as Aharonovich & Benayahu (2011), found the utility of ultrastructure as a character for distinguishing species.This is observed at a much higher level of detail, focusing on the depositional form of calcium carbonate crystals.
The present study offers a new and comprehensive morphological description of Renilla muelleri sclerites, accompanied by an analysis of their size variations among colonies and within different zones of colonies.Furthermore, it presents evidence suggesting that both the external microscopic morphology and biometry of the sclerites exhibit variability between colonies and within colonies, challenging their reliability as taxonomic characters at a specific level.To further validate these observations, future studies employing the same methodology on other congeneric species are recommended, allowing for comparative analyses.Finally, future studies on Renilla muelleri should aim to explore explanations for the patterns of variation in sclerite sizes and their relationships with colony size, as well as environmental variables such as water column pressure, currents, and predation.ACKNOWLEDGMENTS: This research is part of FMG's degree work.We express gratitude to Agustin Schiariti, Gabriel Genzano, and Laura Ferrero for their valuable comments that enhanced this work.We also want to thank two anonymous reviewers for their constructive contributions.Finally, we thanks Jose Luis Martin for his help in the final presentation of the plots.

Figure 2 .
Figure 2. Study area and sampling stations.The triangle corresponds to the place of the Undine ('U') campaign and the circle to the place of the Holmberg ('H') campaign.

Figure 3 .
Figure 3. Kernel density plot of the length (µm) of all sclerite types according to the zone of Renilla muelleri colonies.

Figure 4 .
Figure 4. Kernel density plot of the length (µm) of (a) small and (b) large sclerites according to the zone of the Renilla muelleri colony.

Figure 5 .
Figure 5. Stratified boxplot for the length (µm) of each type of sclerite according to the zone of Renilla muelleri colonies.
Figure 8.(a) and (b): axis with rounded ends of the same width as the central region.Sa: small sclerite with regular edges and complete linear longitudinal grooves.Sb: small sclerite with irregular edges and incomplete spiral grooves.La: large sclerite with irregular edges and incomplete linear longitudinal grooves, Lb: large sclerite with regular edges and complete longitudinal spiral grooves.(c) and (d): Axis with acute ends, the central region is the widest, Sa: small sclerite with irregular edges and incomplete longitudinal spiral grooves.Sb: small sclerite with regular edges and incomplete linear longitudinal grooves.La: large sclerite with irregular edges, and complete longitudinal spiral grooves.Lb: large sclerite with regular edges and incomplete linear longitudinal grooves.
: FMG, AG: Methodology, Software, Data curation, Formal analysis, FMG: Writing -original draft, Visualization, Investigation; AG: Supervision; FMG, FHA, AG: Conceptualization, Writing -review & editing.All the authors actively participated in the discussion of the results, they reviewed and approved the final version of the paper.CONFLICT OF INTEREST: Authors declare there are no conflicts of interest.FUNDING INFORMATION: The study received funding from grant PIBAA No. 0435 (CONICET) and 1037/21 (UNMdP) awarded to AG.

Table 2 .
ANOVA of the models for length differences between individuals.* Denotes the selected model.AIC: Akaike's Information Criterion, with the lowest value indicating the best-fitted model.

Table 4 .
Renilla muelleri.Rachis, sclerite large.Estimated values and confidence intervals derived from GLM, along with the p-value of the t-test for β₁ of the model.
* Indicates significant differences for α = 0.05.The p-value for individual 1H is not displayed, as it was employed for comparisons."H" denotes colonies from the Holmberg campaign, and "U" corresponds to colonies from the Undine campaign.

Table 5 .
Renilla muelleri.Peduncle, sclerite small.Estimated values and confidence intervals derived from GLM, along with the p-value of the t-test for β₁ of the model.* Indicates significant differences for α = 0.05.The p-value for individual 1H is not displayed, as it was employed for comparisons."H" denotes colonies from the Holmberg campaign, and "U" corresponds to colonies from the Undine campaign.

Table 6 .
Renilla muelleri.Peduncle, sclerite large.Estimated values and confidence intervals derived from GLM, along with the p-value of the t-test for β₁ of the model.
* Indicates significant differences for α = 0.05.The p-value for individual 1H is not displayed, as it was employed for comparisons."H" denotes colonies from the Holmberg campaign, and "U" corresponds to colonies from the Undine campaign.

Table 7 .
ANOVA of the models for differences in Length between zones.* Denotes the selected model.** Indicates significant differences for α = 0.05.AIC: Akaike's Information Criterion, with the lowest value corresponding to the best-fitted model.

Table 8 .
Estimate and confidence intervals calculated from the GLMM model.P-value of a t-test of the β i of the peduncle vs. rachis model.