The regeneration levels of the two species from the HSS included in the analysis were influenced significantly by most of the forest structure and physical environmental factors. Regeneration levels of both species were influenced significantly by TD and BA (positively and negatively, re- spectively). PAR and SDb had significant negative effects on the regeneration levels of both species (Table 8). RC had no significant influence on the regeneration of S. taxoides, but this factor positively affected the regeneration of Alph- onsea elliptica. All other factors influenced the regeneration of both species (Table 8). The regeneration of the dominant species in the HSS tended to be influenced by dense, small-stemmed trees under low light, and by soft soil texture. Restricted tree regeneration in this site was de- termined by most environmental factors. The densely packed stems of adult trees and the shade cast by their canopy strongly reduced illumination. Species growing at such sites are likely shade demanding, rather than shade tolerant (Yangetal.2011).Thelowlightregimeoftheforestunder- story favors seedling growth in more shade-tolerant species (Palomaki et al. 2006; Read et al. 2015). A steep light gra- dient without canopy gaps or a moisture gradient across a uniform topography may affect juvenile survival (Aiba and Nakashizuka 2007). Soft soil texture likely regulates both seedling establishment and survival (Uselman et al. 2015).
The regeneration levels of the three species from the SHS included in the analysis were influenced more by physical environmental than by forest structure factors. Only one forest structure factor had a significant effect. The effects of the physical and forest structure factors varied among the three species (Table 8). The regeneration values of two of the three SHS species were affected negatively by BA. Only one species was affected positively by TD (Table 8). SMC had significant positive effects on the regeneration levels of Memecylon caeruleum and Diospyros castanea, which var- ied among other physical factor levels and factors related to forest structure. On the steep slopes of limestone hills, rain- water penetrates rapidly through the thin soil layer and con- nected fissures in the bedrock layer to layers below the root- ing zone (Butscher and Huggenberger 2009). However, dense bedrock in karst landforms enhances the surface soil moisture in the epikarst, thereby contributing positively to the colonization process of tree species and vegetation de- velopment in the area (Li et al. 2014). Soil moisture likely regulates both seedling establishment and survival, and is a key factor in the initial process of seedling recruitment (Khurana and Singh 2001; McLaren and McDonald 2003).
Lamthai Asanok and Dokrak Marod
Trejo-torres and Ackerman (2002) indicated that the soil humidity regime, which is related to topography, was more important than geography in promoting floristic links among limestone vegetation communities of the Caribbean. Dry communities on single islands had stronger floristic af- finities with dry communities on other islands than with more humid vegetation on the same islands (Felfili et al. 2007). The regeneration values of Phyllanthus mirabilis, one of three dominant species in the SHS, had significant positive relationships with RC and PAR, but were related negatively to SDb (Table 8). This suggests that this species established well in limestone outcroppings, illuminated sites, and thin soil sites. Tree species that regenerate on rocky out- crops usually have specialized root systems that are able to thread through the rocky substratum and extract nutrients fromthethinsoil/humuslayer(Denketal.2014;Duarteet al. 2015). The roots commonly grow deep into the rock fis- sures of the epikarst substratum (Huang et al. 2011). This species also occurred at illuminated sites, suggesting that it requires high light conditions and has rapid growth rates (Carreño-Rocabado et al. 2012; Lusk et al. 2013).
The regeneration levels of the two species from the HCS included in the analysis were influenced more by physical environmental than by forest structure factors (Table 8). The regeneration levels of both species included in the anal- ysis had significant positive relationships with RC and PAR. The regeneration of Ficus macleilandii was determined by BA, RC, and PAR (positive or negative effects). The re- generation value of Dracaena cochinchinensis was affected positively by RC and PAR, but negatively by SDb (Table 8). Physical factors, especially RC and PAR, had greater significant positive effects on the regeneration of the domi- nant species at the HCS than did factors related to forest structure. This site was relatively dry due to the very shal- low soils and relatively high summer temperatures caused by intense solar radiation, which exceeded the levels in sur- rounding areas. The cliffs are protected as special habitats, but access is allowed to interesting v