Species richness and phylogenetic diversity of seed plants across vegetation zones of Mount Kenya, East Africa

Abstract Mount Kenya is of ecological importance in tropical east Africa due to the dramatic gradient in vegetation types that can be observed from low to high elevation zones. However, species richness and phylogenetic diversity of this mountain have not been well studied. Here, we surveyed distribution patterns for a total of 1,335 seed plants of this mountain and calculated species richness and phylogenetic diversity across seven vegetation zones. We also measured phylogenetic structure using the net relatedness index (NRI) and the nearest species index (NTI). Our results show that lower montane wet forest has the highest level of species richness, density, and phylogenetic diversity of woody plants, while lower montane dry forest has the highest level of species richness, density, and phylogenetic diversity in herbaceous plants. In total plants, NRI and NTI of four forest zones were smaller than three alpine zones. In woody plants, lower montane wet forest and upper montane forest have overdispersed phylogenetic structures. In herbaceous plants, NRI of Afro‐alpine zone and nival zone are smaller than those of bamboo zone, upper montane forest, and heath zone. We suggest that compared to open dry forest, humid forest has fewer herbaceous plants because of the closed canopy of woody plants. Woody plants may have climate‐dominated niches, whereas herbaceous plants may have edaphic and microhabitat‐dominated niches. We also proposed lower and upper montane forests with high species richness or overdispersed phylogenetic structures as the priority areas in conservation of Mount Kenya and other high mountains in the Eastern Afro‐montane biodiversity hotspot regions.

species in an area; Faith, 1992) is a biodiversity index which quantifies the combined phenotypic or genetic diversity across the species (Cadotte & Davies, 2016;Davies et al., 2007;Purvis & Hector, 2000). Phylogenetic diversity reflects the underlying evolutionary history represented by a set of taxa, and it has become a subject of interest to both ecologists seeking to understand the influence of evolutionary history on species abundance and interactions, and to conservation biologists wishing to prioritize evolutionary history for conservation purposes (Forest et al., 2007;Li et al., 2015;Sechrest et al., 2002;Tucker et al., 2017). There is a positive relationship between phylogenetic diversity and species richness (Kluge & Kessler, 2011;Sax et al., 2007), especially along an elevational gradient where changes in species richness are apparent (Li et al., 2015).
The relative similarity of species and phylogenetic diversity across space is dependent on the underlying environmental conditions and biogeographical history. It could be that species richness and phylogenetic diversity are highly congruent due to random or even selection of taxa across a phylogeny (Rodrigues, Brooks, & Gaston, 2005). Conversely, highly nonrandom patterns, for example, phylogenetic signals in environmental tolerances or localized speciation events, could create incongruent taxonomic and phylogenetic diversity patterns (Devictor et al., 2010;Tucker & Cadotte, 2013). Given these complexities, both taxonomic and phylogenetic diversity should be evaluated for conservation. For example, Li et al. (2015) found that compared to other communities, the evergreen broad-leaved forests in Dulong Valley in China had the highest levels of species richness and phylogenetic diversity, as well as an overdispersed phylogenetic structure, and suggest that communities with high species richness or an overdispersed phylogenetic signal should be the focus for biodiversity conservation, as these areas may help maximize the potential of local flora to respond to future global change.
When evaluating diversity patterns, it is also important to compare and contrast species belonging to different ecological guilds. This is due to the fact that differences in life history traits or mechanisms of resource uptake can lead to highly disparate responses to outside inputs such as climate. For plants, the distinction between woody and herbaceous growth forms is probably the most profound contrast in terrestrial ecosystems . Herbaceous and woody taxa are believed to be differentially influenced by environmental factors such as precipitation and temperature (Whittaker, 1965), which are nonrandomly associated with the evolutionary history of these taxa and their traits (Díaz et al., 2016). Thus, it may be that factors resulting in variation in species richness and phylogenetic diversity might differ between woody and herbaceous plants, and comparisons between these communities may provide better insight into the factors influencing the distribution of total plant biodiversity (Bhattarai & Vetaas, 2003).
Here, we analyze changes in taxonomic and phylogenetic diversity across different vegetation zones of Mount Kenya, which is the largest ancient extinct volcano in the Great Rift Valley area, and the second highest peak in Africa (Speck, 1982). It constitutes an important reservoir for plant diversity, including a substantial number of endemic and endangered species. As the publishing of the first checklist of about 140 plant species by Hooker and Oliver in 1885, numerous studies have studied plant and vegetation diversity in Mount Kenya (e.g., Bussmann, 1994;Fries & Fries, 1948;Niemelä & Pellikka, 2004;Young & Peacock, 1985). However, plant taxonomic  Niemelä & Pellikka, 2004). (c) The vegetation zones of Mount Kenya from northwest to southeast in lateral view (adapted from Coe, 1967 andVECEA Team, 2012) and phylogenetic diversity of Mount Kenya have not been thoroughly analyzed, and we lack critical knowledge on the evolutionary dimension of the biodiversity in this region.
The aim of this study was to quantify the relationship between species richness and phylogenetic diversity and to explore community phylogenetic structure across vegetation zones of Mount Kenya, Kenya. We also aim to compare the diversity patterns between woody and herbaceous plants in order to a get better insight into the factors contributing to observed patterns in diversity.

| Data sources
We compiled a comprehensive checklist of seed plants based on data from various scientific expeditions to Mount Kenya since the 1900s.
These data sources include published floras and field guides such as Flora of Tropical East Africa (FTEA editors, 1952(FTEA editors, -2012, Upland Kenya Wild Flowers and Ferns (Agnew, 2013), Wild Flowers of East Africa (Blundell, 1987) and Kenya Trees Shrubs and Lianas (Beentje, 1994), data of specimens from East African Herbarium, Nairobi, Kenya (EA) and Global Biodiversity Information Facility (GBIF, https://www. gbif.org/), and data from our own scientific expedition from 2009 to 2016 (specimens were stored at Herbarium of Wuhan Botanical Garden, Wuhan, China, HIB). Species were assigned to vegetation zones according to the sampling locations and habitat descriptions described in the monographs. Growth forms of species were classified as either woody or herbaceous plants.

| Taxonomic metrics
To eliminate the area effect on species richness in zones of different sizes, species density (D) for each zone was calculated based on the following equation (Li et al., 2015;Vetaas & Grytnes, 2002): where S is number of species in each zone and A is the area of each zone.

| Phylogeny construction
Our primary goal was to calculate phylogenetic distance metrics and we first constructed a phylogenetic tree for all the seed plants

| Phylogeny metrics
We calculate two diversity measures PD and SES_PD, for total, woody, and herbaceous plants of each zones. Faith's phylogenetic diversity (PD) is widely used in several conservation studies, although it is positively correlated with species richness (Li et al., 2015). A null model to standardize PD measurements and standard effect size phylogenetic diversity (SES_PD) was also calculated by for each analysis to assess the statistical significance of the observed patterns. A two-tailed significance test was used to assess whether these NRI/NTI results differed significantly from zero.
Consequently, positive NRI/NTI values indicate phylogenetic clustering that species are more closely related than expected, and negative NRI/NTI values indicate phylogenetic overdispersion and species in communities are more distantly related than expected (Webb et al., 2002). All the phylogenetic analyses were performed in R 3.3.3 software (R Core Team, 2017) with the picante package (Kembel et al., 2010). herbaceous climbers and 842 herbs) ( Figure 3). As expected, species richness, density, and phylogenetic diversity were found to be the highest in low montane dry forest (LMDF) and the lowest in nivial zone (NZ), while the SES_PD was found to be the highest in low montane wet forest (LMWF) and the lowest in Afro-alpine zone (AZ) ( Table 2).

| Species richness, density, and phylogenetic diversity of seed plants
There were notable differences between the diversity of herba- for both woody and herbaceous plants, while the lowest SES_PD was both found in AZ (Table 2).

| Phylogenetic structure of total, woody, and herbaceous plants
When we examined differences between forest zones (LMWF, LMDF, BZ, and UMF) and alpine zones (HZ, AZ, and NZ), both NRI and NTI showed substantially increased pattern while NRI is not significant (P = 0.191) and NTI is significant (P < 0.05; Figure 4). NRI and NTI showed different patterns in total, woody, and herbaceous plants ( Figure 5). For total plants, NRI of LMDF and NTI of LMWF

| D ISCUSS I ON
The species richness of woody and herbaceous plants of forests is expected to be affected by environmental variation, forest canopy, and anthropogenic disturbances (Clinton, 2003;Schmitt, Denich, Demissew, Friis, & Boehmer, 2010;Zhang, Huang, Wang, Liu, & Du, 2016). Woody plants are probably more strongly influenced by largescale environmental variations than herbaceous plants, such as the changes of moisture, temperature, or altitude (Schmitt et al., 2010).
That is to say, the difference in humidity between the southeast and northwest slopes of Mount Kenya has a stronger influence on species richness of woody plants, and lead the LMWF to have the highest level of species richness of woody plants. The density of forest canopy could influence the light intensity, moisture, and temperature in the understory (Clinton, 2003;Whittaker, 1960). There are many big trees found in LMWF, such as Newtonia buchananii (Fabaceae), Ocotea usambarensis (Lauraceae), Vitex keniensis (Lamiaceae), Tabernaemontana stapfiana (Rubiaceae), and Zanthoxylum gilletii (Rutaceae) ( Table 1). These tall trees as well as other small trees, shrubs, or lianas form highest level of species density of wood plants, which blocks sunshine, leading to stunted development of understory herbaceous plants (Table 2). In contrast, LMDF is dominated by small trees and shrubs, as well as few large trees, such as Juniperus procera (Cupressaceae) and Olea europaea (Oleaceae), and this leads to better development of understory herbaceous plants due to light availability. In addition, the areas with moderate impacts may have greater environmental heterogeneity, which could provide greater opportunities for plants to growth (Pausas & Austin, 2001;Zhang et al., 2016). The degree of human disturbance on the northwest slope of Mount Kenya is obviously greater than that in the southeast slope, for example, the two famous mountaineering Interestingly, the NRI of herbaceous plants of AZ and NZ was smaller than those of BZ, UMF, and HZ, and a similar pattern could be found in NRI of total plants ( Figure 5). This phenomenon was also found in areas above 5,500 m asl. in Hengduan Mountains, China, and this can be attributed to the fact that in the upper elevation zones, the plants were sparsely distributed and this lead to a decrease in interspecific competition (Li, Zhu, Niu, & Sun, 2013). While, in woody plants, there are obvious increasing trends of NRI and NTI as the zones transition from UMF to NZ ( Figure 5). Combining the phylogenetic structure patterns of woody and herbaceous plants, we strongly support that, large woody plants have climate-dominated niches, whereas herbaceous plants have edaphic and microhabitatdominated niches, and the temperature or climate filtering process presumably has played a greater role in structuring species into local communities for woody plants than for herbaceous plants (Qian, Jin, & Ricklefs, 2017;Ricklefs & Latham, 1992).
TA B L E 2 Species richness (SR), density (D), phylogenetic diversity (PD), and standard effect size phylogenetic diversity (SES_PD) of total, woody, and herbaceous plants among different vegetation zones of Mount Kenya  Mount Kenya, the upper montane forest is also called Cloud forest or Moist forest, although rainfall in this region is lower than in the montane rainforest zone, evaporation is also lower, and frequent heavy mists contribute to the humidity (Niemelä & Pellikka, 2004).

It has different characteristics and includes tree species such as
Hagenia abyssinica (Rosaceae), Hypericum revolutum (Hypericaceae), and Juniperus procera (Cupressaceae) (Lange et al., 1997;Niemelä & Pellikka, 2004). Recently, several studies explicitly compared biodiversity conservation prioritizations based on both species richness and phylogenetic diversity criteria to assess the efficacy of each approach (Forest et al., 2007;Kraft, Baldwin, & Ackerly, 2010;Li et al., 2015;Vandergast, Bohonak, Hathaway, Boys, & Fisher, 2008). These authors suggested that biodiversity conservation is maximized by the inclusion of communities and zones with overdispersed phylogenetic structure, because such natural areas include phylogenetically distantly related lineages (Li et al., 2015). Other have suggested that proportional endemism is a more important consideration than phylogenetic diversity in developing conservation priorities (Brewer, 2017), but we were not able to assess proportional endemism of each vegetation of Mount Kenya in the current study. In this case, the two lower montane forest zones

(LMWF and LMDF) and the upper montane forest (UMF) of Mount
Kenya should be given as much attention in conservation as the alpine zones because these forest zones are the most phylogenetically diverse and also have the highest species richness which is the key to the maintenance of biodiversity (Vanleeuve et al., 2003).
The mountain chains of eastern Africa, which extend from the Ethiopian mountains, through east African nations and southwards to Mozambique, are formed as the result of an active continental rift (Chorowicz, 2005). It contains the Eastern Afromontane biodiversity hotspot (EABH), which is the one out of eight biodiversity hotspots of the Afromadagascan region (Mittermeier, Turner, Larsen, Brooks, & Gascon, 2011 Our findings have a profound effect on biodiversity conservation across the EABH as the vegetation types in these mountains are similar to those of Mount Kenya (Bussmann, 2006). It is important that the lower and upper montane forests of EABH should be given as much attention in conservation just like Mount Kenya, for they also have the highest species richness and most diverse phylogenetic lineages, in addition to having the highest evolutionary potential (Faith, 1992;Forest et al., 2007;Li et al., 2015).

ACK N OWLED G M ENTS
We would like to thank Z. Zhong from Wuhan Botanical Garden,

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N
Y.Z. and S.C. conceived and wrote the manuscript. Y.Z. and G.H. provided and analyzed the data. G.M., X.Y., and Q.W. provided the idea.
All authors reviewed the manuscript.

DATA ACCE SS I B I LIT Y
Data are available via the Dryad Digital Repository: https://doi. org/10.5061/dryad.7sj4t8h.