The historical biogeography and conservation value of taxonomic distinctness: The case of ferns flora of the Gibraltar Arc

The pteridofloras of nine locations in the Gibraltar Arc were analyzed using a taxonomic distinctness index. We found that the index could be a proxy of historical biogeography of the pteridofloras from this area. Moreover, the value of the taxonomic distinctness index of the different locations showed relevant relationships with certain geographic variables. Finally, we hypothesize about the value of the information derived from taxonomic distinctness index for conservation of the pteridoflora in the Gibraltar Arc.


Introduction
The fern flora of the Gibraltar Arc, i.e., the geological region corresponding to an arcuate orogen surrounding the Alboran Sea, extending from the Iberian Peninsula to Africa, has been extensively studied recently by Salvo Tierra et al. (2020). These authors suggested that the factors influencing diversity in fern floras of nine different mountain areas is so relevant that they should be included in the management of the territory. They concluded that the altitudinal gradient and the amplitude of the dry period have determined the distribution of pteridophytes in the Gibraltar Arc, a distribution in which the longitudinal gradient predominates over the latitudinal gradient. In particular, the oldest fern floras (with greater diversity) are located at the western ends of the arc on both continents. Five pterido-geographic units were detected, two of them spanning S Europe and N Af-rica, and eleven chorotypes. Finally, considering the diversity, age, singularity, and degree of threat to its pteridoflora, it was proposed that the Gibraltar Arc could be considered as a 'Sanctuary for ferns', an increasingly widespread conservation concept.
Taxonomic structure (based on the number of taxa included in higher levels of the taxonomic hierarchy) has been widely used in ecology to measure impacts on the diversity of ecosystems (for example, Clarke & Warwick 1998, 1999, 2001, Tucker et al. 2017, Pérez Hernández 2019. Taxonomic structure shows a latitudinal gradient and is maximum at the Equator (Krug et al. 2008). Among taxonomic richness estimators, taxonomic distinctness (TD; Pérez Hernández 2019, and references therein) is particularly significant because it appears to be less influenced by sample size than are species richness or Shannon species diversity. Moreover, TD may be a more sensitive univariate index of community perturbation than spe-cies diversity (Pérez Hernández 2019, and references therein). Furthermore, the statistical signifi cance of the TD index can be tested (Clarke & Warwick 2001, Pérez Hernández 2019. Clarke & Warwick (1998) proposed the ∆+ index (i.e., Delta-plus) to estimate taxonomic distinctness. This index measures the overall mean taxonomic path length between any two randomly chosen taxa. Because the index has no dependence on sampling effort, the ∆+ fi gures computed from different sources of data can be compared across studies with differing and uncontrolled degrees of sampling effort.
The main aim of the current note is to analyze the TD of fern fl oras from the Gibraltar Arc. The goal is to assess the differences in biogeographical composition on opposite shores of the Alboran Sea.

Data origin
The determination of the OGUs (operative geographical units) was carried out according to criteria of geophysical homogeneity and in correspondence with the biogeographic units, according to Salvo Tierra et al. (2020). Nine mountain areas located in the Gibraltar Arc were selected, one for each biogeographical sector; the areas include the entire range of thermoand ombrotypes found in these regions (Rivas Martinez 2008): Gata, Nevada, Tejeda, Ronda, Aljibe, Tingitania, Chaouen, Gurugú and Tlemcen (Fig. 1). The basic matrix of data is showed in the supplementary data (metadata available in Salvo Tierra et al., 2020).

Data Analysis
The ∆+ index is computed according to the equation: where s is the number of species observed, and w ij is the weight or distinctiveness value given to each taxonomic branch of the hierarchical classifi cation, from species i to the fi rst node in common with species j. Thus, w ij = 0 if i and j are the same genus, w ij = 1 if they are the same family.
In addition, in a second step, linkage cluster analysis of the ∆+ index per locality was performed using a multivariate agglomerative method. Thus, we used the Bray-Curtis distance to generate the similarity matrix, using the free PAST version 4.02 software (Hammer et al. 2001). We chose Bray-Curtis distance to generate the similarity matrix, because it is especially recommended for standardized data (Yoshioka 2008) (Fig. 2).
For each OGU, the values of latitude, longitude, maximum altitude, TD and TD range values (the difference between the upper and lower limits), specifi c richness, index m/t (Ito 1972(Ito , 1978 and degree of ploidy, as well as the percentages of the presence of each of the generalized tracks detected in the study area, were included in the analysis (Salvo Tierra et al. 2020).

Results and Discussion
The ∆+ index for each region (i.e., Tingitania, Chauen, Gurugú, Tlemcen, Aljibe, Ronda, Tejeda, Nevada, and Gata) was estimated. We did not observe signifi cant differences according to the criteria of the index. However, we observed three different and coherent similar groups according their ∆+ values ( Fig. 1). Because the signifi cance of the index was developed under other criteria (in relation to the number of sampling stations, and the number of species) (Clarke & Warwick 1998, 1999, 2001), we consider that they may not be useful here.
From Table 1 we conclude: 1) There is a strong inverse correlation between the value of TD and the maximum altitude (the lower the altitude, the greater the distinctness), as well as with the geographical longitude (decreasing from west to east) and with the index m/t. 2) From this last analysis, it can be inferred that there is a correspondence between higher values of TD and the older pteridofl oras (the proportion of taxa with trilete spores = primitive character, is high compared to those with monolete spores = advanced character). 3) This same observation is corroborated on the one hand by analysing the correspondence between TD and the degree of ploidy (the lowest values of the degree of ploidy correspond to the most primitive pteridofl oras), and on the other by observing that the lowest values of TD are correlated with those pteridofl oras in which the generalized tracks of wider distributions (heterochoric and circumboreal) have a greater weight. 4) Consequently, the highest values of TD correspond to the most relict pteridofl oras, although they arose from different historical events, and are scarcely 'contaminated' by species with wider geographical ranges that were incorporated into the territory during subsequent events. 5) Finally, TD has a complementary signifi cance in the defi nition of priority conservation areas. In this sense, the lower values in the range of the upper and lower limits of TD for Aljibe and Nevada indicate a more ancestral and consolidated flora, which would support a high priority as regions for the conservation of pteridophytes. This is confirmed when applying the Vane-Wright et al. (1993) these two units are defined in the first cycle on the application method.