Abstract
A large number of small-sized samples invariably shows that woody species are absent from forest soil seed banks, leading to a large discrepancy with the seedling bank on the forest floor. We ask: 1) Does this conventional sampling strategy limit the detection of seeds of woody species? 2) Are large sample areas and sample sizes needed for higher recovery of seeds of woody species? We collected 100 samples that were 10 cm (length) ×10 cm (width) ×10 cm (depth), referred to as larger number of small-sized samples (LNSS) in a 1 ha forest plot, and placed them to germinate in a greenhouse, and collected 30 samples that were 1 m×1 m×10 cm, referred to as small number of large-sized samples (SNLS) and placed them (10 each) in a nearby secondary forest, shrub land and grass land. Only 15.7% of woody plant species of the forest stand were detected by the 100 LNSS, contrasting with 22.9%, 37.3% and 20.5% woody plant species being detected by SNLS in the secondary forest, shrub land and grassland, respectively. The increased number of species vs. sampled areas confirmed power-law relationships for forest stand, the LNSS and SNLS at all three recipient sites. Our results, although based on one forest, indicate that conventional LNSS did not yield a high percentage of detection for woody species, but SNLS strategy yielded a higher percentage of detection for woody species in the seed bank if samples were exposed to a better field germination environment. A 4 m2 minimum sample area derived from power equations is larger than the sampled area in most studies in the literature. Increased sample size also is needed to obtain an increased sample area if the number of samples is to remain relatively low.
Highlights
The presence of viable seeds in the soil has been extensively documented [1,2,3], and seeds in the soil may include those dispersed directly from existing plants at the site, persisting from previous vegetation, and arriving via secondary dispersal from other sites [4,5,6]
Only 13 species from 11 genera of 9 families were identified from large number of small samples (LNSS) in the greenhouse, which only accounted for 15.7% of woody species in the forest stand (Table 1)
Even with 100 small-sized samples only 13 woody species were found by the LNSS method
Summary
The presence of viable seeds in the soil has been extensively documented [1,2,3], and seeds in the soil may include those dispersed directly from existing plants at the site, persisting from previous vegetation, and arriving via secondary dispersal from other sites [4,5,6]. The precision (narrow confidence interval) of seed-number estimates may be improved by taking a large number of small samples (LNSS) [11]. Hopfensperger [15] reviewed 108 articles published between 1945 and 2006 on the similarity between above and belowground species composition and found lower similarities in forests than in grasslands and in wetlands Another comprehensive review from Europe showed low similarities between forest-soil seed banks and standing vegetation [16]. On the forest floor, many species produce seedlings under a range of environmental conditions This large discrepancy between lack of seeds in the soil seed bank studies and presence of seedlings limits valid interpretations of the soil seed bank and forest stand relationships, and it could lead to misjudgment of restoration potentials in degraded and disturbed forest sites
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
