The study was carried out in the moist tropical forests in the north of the Republic of Congo and more particularly within the forest management unit of Loundoungou, granted to the company Olam/Agri. Geographical coordinates of forest management unit of Loundoungou are: 02˚18' - 02˚22'N and 17˚31' - 17˚34'E (Figure 1). The area is relatively flat with average altitudes between 430 m and 530 m. It displays a bimodal distribution of seasonal precipitation. The average annual rainfall and the average temperature are respectively 1729 mm and 25˚C (Bégué, 1967) . The forest management unit of Loundoungou is a semi-deciduous forest, installed on clay soils of the Congolese Cuvette (Gond et al., 2013; Freycon, 2014; Fayolle et al., 2014) . It has been frequently disturbed in the past by traditional human activities (agricultural activities, hunting, etc.) (Oslisly et al., 2013; Morin-Rivat et al., 2014) . It is composed of numerous light-demanding tree species such as: Erythrophleum suaveolens (Guill. & Perr.) Brenan, Celtis spp., Terminalia superba Engl. & Diels, Petersianthus macrocarpus (P. Beauv.) Liben and Triplochiton scleroxylon K. Schum. Some places of the undergrowth are invaded by lianas and giant herbaceous belonging to the Marantaceae and Zingiberaceae families (Gillet, 2013; Gond et al., 2013) .

The soil samples were collected within two plots of 250,000 m² (one plot exploited in 2008 and another in 2021, named P08 and P21, respectively) in September 2021 at the end of the great dry season and the beginning of the short

rainy season. Within each plot of 250,000 m², soil samples were collected in two plots of 20 m × 20 m, installed at the triangular end of each plot. Inward of the plots of 20 m × 20 m, each plot has been divided in four sub-plots of 10 m x 10 m, i.e. a unit area of 100 m² inside which soil samples were collected.

The soil has been collected using a parallelepipedic box in wood measuring 10 cm long × 10 cm wide × 5 cm deep. Sixteen collection points for all four plots were established in the center of each sub-plot, i.e. eighty composite samples for all soil layers combined.

Five superimposed soil layers were collected per point: 0 - 5 cm, 5 - 10 cm, 10 - 15 cm, 15 - 20 cm and 20 - 25 cm. Each sample was composite, that is to say from of mixture of three unit taking realized on the summits of an equilateral triangle of 1 m side, i.e. a volume per layer and per point of 1500 cm³ (Perera, 2005; Daïnou et al., 2011; Douh et al., 2018a) . The three unit taking from each sampling point were putting in a plastic bag, which contained information: type of the plot and sub-plots, number of collection point and the level of taking of the soil layer corresponding either to the plot exploited in 2008 (P08) and/or plot exploited in 2021 (P21). A total volume of 240,000 cm³ has been obtained for the two plots (P08 and P21), i.e. 2 × 120,000 cm³ per plot, over a total sampling area of 0.48 ha (2 × 0.24 ha per plot).

Two weeks after collecting the soil samples, they were dispatched to Brazzaville and more precisely to the departmental nursery of SNR (Service National de Reboisement) for germination tests.

The number of viable seeds in the soil was estimated by the germination method (Simpson et al., 1989; Sousa et al., 2017; Douh et al., 2018a) . The soil samples were sieved with a 2 mm mesh size item in order to eliminate crude pieces of plants and coarse mineral elements (e.g. leaves, pieces of wood and stones). A total of two blank trays (containing only sterilized substrate) per plot have been used to verify eventual contamination of the forest soil samples; no germination was noted from the blank samples at the end of the observation period. In fact, the germination was carried out in a nursery at the SNR (Service National de Reboisement) site in Brazzaville in a nonforest zone in order to prevent contamination of the samples by seeds of the surrounding vegetation.

The shadehouse’s relative light intensity was about 30% - 40% of full sunlight. Each soil sample was spread out over a sterilized (by heat) sand substrate in a way that the average depth of the sample should not exceed 1 cm (Hall & Swaine, 1980; Douh et al., 2018a) . Watering was almost daily. The monitoring of germination was performed twice a week (every 3 - 4 days). When their development stage allowed it, the seedlings were removed and transplanted individually in polyethylene bags in order to favour their growth and identification. As all the germinations took place over the course of the first 17 weeks, the experiment was stopped after 20 weeks (i.e. 5 months of experimentation). Voucher specimens were collected and species were identified by botanists (Gilbert Nsongola, Isaac Dzombo). We followed the taxonomy of the Geneva Herbaria Catalogue (http://www.ville-ge.ch/musinfo/bd/cjb/chg).

To assess the effect of soil samples collection depth on seeds germination, the Generalized Linear Model (GLM) in terms of the explanatory variable “soil collection depth”, followed by a classification according to Student-Newman-Keurls (SNK) between the average number of germinated seeds per soil layer within the two plots (P08 and P21) were realized using SPSS version 22.0 software (Nelder & Wedderburn, 1972) . To identify the characteristics of the species observed in the soil seed bank, we classified species in terms of light-requirements and dispersal syndrome. We used the regeneration guilds of tropical forest tree species defined by Hawthorne (1995) and the dispersal syndromes of seed defined by Armesto & Rozzi (1989) and Howe & Smallwood (1982) . The herbaceous species, the liana species and the undetermined species were not taken into account. The following parameters were used to describe the abundance of seeds of two plots (P08 and P21): the absolute density AD (seeds/m²), the relative density RD (%, number of seeds of a given species/the total number of seeds for all species), the relative frequency RF (%, proportion of samples containing the given species) and the species Importance Value Index (IVI) computed as the sum of RD and RF (Mueller-Dombois & Ellenberg, 1974; Borges & Engel, 1993; Butler & Chazdon, 1998; Martins & Engel, 2007; Daïnou et al., 2011) . To identify the indicator species of each plot based on seedlings from the soil bank, we computed the “Indicator value index” (INDVAL), using the labdsv package implemented in the R environment (Roberts, 2012) . Significance was set at p < 0.05. This index described by Dufrêne & Legendre (1997) is defined as follows:

INDVAL = A i j × B i j × 100

A i j = N i n d i v i d u a l s i j / N i n d i v i d u a l s i (1)

B i j = N s i t e s i j / N s i t e s j

INDVAL = the Indicator Value of species in site group j.

A i j , is a measure of specificity (based on the abundance of species i);

N i n d i v i d u a l s i j , is the mean number of individuals of species i in the sites of group j;

N i n d i v i d u a l s i , is the number of individuals of species i in all groups;

B i j , is a measure of fidelity (based on incidence of species i);

N s i t e s i j , is the number of sites in the group j where species i is present;

N s i t e s j , is the total number of sites in that group. Here, there were two groups constituted by the two plots, P08 and P21. The undetermined species were excluded from the computation.

In terms of species diversity, we first computed the observed species richness, Sobs. But as Sobs is very dependent on the sampling effort and is considered as an unreliable indicator of the total species richness (Walther et al., 2005) , we also computed estimated species richness using two of the most recommended estimators: the bias-corrected Chao2, Schao2 (based on incidence) and Jackknife1, Sjack1 (based on abundance) (Walther & Morand, 1998; Walther & Martin, 2001; Chiarucci et al., 2003; Dove & Cribb, 2006) , using the program EstimateS 9.1.0 (Colwell, 2013) . Finally, the specificity of layers in terms of species found exclusively in each layer was determined by comparing these “exclusive” species with the total number of species found in the layers (%Sexcl).

Rarefaction curves were derived from the observed and estimated species richness to evaluate the representativeness of our sampling effort. Finally, within both plots, viable seeds density per soil layer was computed cumulating the data of each point per soil layer. The differences of mean seed densities between the two plots have been tested using a parametric Student’s T test at the 5% threshold, the normality of the data having been verified by the Shapiro-Wilks and Kolmogorov-Smirnov tests (Nelder & Wedderburn, 1972; McCullagh & Nelder, 1993) .

Nevertheless, to compare the floristic composition of soil seed bank between the two plots (P08 and P21), we performed a nonmetric multidimensional scaling (NMDS) based on seed abundance data (8 collection points per plot). NMDS, applied with the R package MASS (Ripley et al., 2017) , makes no assumptions about the data (Faith et al., 1987) , and is considered among suitable methods for graphical representation of floristic ordination (Clarke, 1993; Glèlè Kakaï et al., 2016) . The undetermined species were not taken into account in the ordination. The samples that did not provide any germination were also excluded from the analysis. Shannon’s (H’) and Pielou’s Equitability (E) indices were respectively used to evaluate the specific diversity and the equitable distribution of individuals from the soil seed bank between the two plots and the soil layers (Cordonnier et al., 2012; Djego et al., 2012) . In addition, to study the variations of diversity indices between different soil layers, we used the nonparametric Kruskall-Wallis test with post-hoc pairwise multiple comparisons between the means for a probability value of p-value < 0.05.

Table 1 demonstrates that the explanatory variable “soil sampling depth” had a very highly significant effect (p-value = 0.000) on the germination of the soil layers of the plot exploited in 2008 (P08). On the other side, it did not have a significant effect (p-value = 0.085) on the germination of the soil layers of the plot exploited in 2021 (P21) (Table 1).

Nevertheless, in the P08, two classes of average germinations per sample were identified on the basis of the classification SNK (Student-Newman-Keuls) (Table 2).

The first class, composed of the 15 - 20 cm, 10 - 15 cm and 20 - 25 cm layers, displays relatively low average germinations per sample (Table 2). The second class, composed of the 0 - 5 cm layer, stood out others soil layers by a strong average of germinations per sample (Table 2).

However, in the plot exploited in 2021, the classification SNK (Student-Newman-Keuls), demonstrates that there are no significant differences between the average germinations per sample on the soil layers (Table 2).

We recorded 347 seedlings in the plot exploited in 2008 (P08) and 418 seedlings in the plot exploited in 2021 (P21), with respective average densities of 1446 seeds/m² and 1742 seeds/m² (Table 3). There were no significant differences between the mean densities of the two plots (T test; F = 2.385; p-value = 0.128). In the P08, the seedlings belonged to 37 species whose 25 herbaceous species (82.71%), 11 trees and shrubs species (17%) and 1 liana species (0.29%). The tree and shrub species identified were Nauclea diderrichii (De Wild. & T. Durand) Merr. (108 seeds/m²), Musanga cecropioides R. Br. (33 seeds/m²), Macaranga barteri Müll.-Arg. (4 seeds/m²), Phyllanthus muellerianus (Kuntze) Exell (4 seeds/m²), Phyllanthus sp. (4 seeds/m²) and Psychotria sp. (4 seeds/m²) (Table 3). The most abundant herbaceous species were Oldenlandia corymbosa L. (425 seeds/m²) and Erigeron sumatrensis Retz (275 seeds/m²) (Table 3). However, in P21, seedlings belonged to 27 species whose 18 herbaceous species (93.78%) and 9 trees and shrubs species (6.22%). Tree and shrub species identified were M. cecropioides (17 seeds/m²), Macaranga spinosa Müll.-Arg. (17 seeds/m²), Nauclea diderrichii (4 seeds/m²) and Ricinodendron heudelotii (Baill.) Pierre ex Heckel (4 seeds/m²). The most abundant herbaceous species were Oldenlandia corymbosa (688 seeds/m²) and Erigeron sumatrensis (608 seeds/m²) (Table 3). Regarding indicator species, P08 displayed a single indicator species, N. diderrichii (IndVal = 0.843%; p-value = 0.004), whereas P21 displayed two indicator species: E. sumatrensis (IndVal = 0.689%; p-value = 0.041) and O. corymbosa (Indval = 0.618%; p-value = 0.032) (Table 3). The observed specific richness (Sobs) of P08 is estimated to 37 species, whereas the specific richness estimators Sjack1 and Schao2 estimate 39 and 40 species respectively (Figure 2). However, that of P21 is estimated to 27 species, whereas the Sjack1 and Schao2 estimators display 31 and 29 species respectively (Figure 2). In all cases, the Sobs curve does not truly approach the asymptote, but increases gradually with the increase

of number samples. Consequently, the sampling effort is relatively considerable in the two plots studied (Figure 2).

Nonetheless, the floristic composition of the soil seed bank of the two plots (P08 and P21), illustrated using the NMDS, demonstrates that there is a spatial structuring of the samples confirming that the species from the seed bank of the soil of the two plots are globally similar (Figure 3).

Concerning the specific richness at the level of the soil layers, the P08 presents a relatively high specific richness compared to the soil layers of the P21 which seems relatively constant at the level of the superficial layers compared to the deepest layers of soil (Table 4). The observed species richness values (Sobs) at the level 0 - 5 cm layers of P08 and P21 are respectively 29 to 14 species. Nonetheless, the Schao2 and Sjack1 estimators displays 36 (33 species) and 29 (15 species), respectively (Table 4). On the other side, the specific richness of the layers 20 - 25 cm is of 13 species for P08 and 16 species for P21. The Schao2 and

Sjack1 estimators respectively estimate 14 (14 species) and 17 (18 species) (Table 4).

Interesting fact, in the 20 - 25 cm layer, we record 7.7% of exclusive species for P08 and 18.8% for P21 (Table 4). The species found exclusively in the 20 - 25 cm layer of P08 was Rubiaceae sp.4. While in P21, the species found exclusively in the 20 - 25 cm layer were Nauclea diderrichii, Rubiaceae sp.5 and Davallia denticulata (Burm. f.) Mett. ex Kuhn. Otherwise, the average seed densities obtained in the 20 - 25 cm layer were 279 ± 125 seeds/m² for P08 and 458 ± 49 seeds/m² for P21, compared to the superficial layers 10 - 15 cm and 15 - 20 cm. What illustrates well the necessity to explore the deeper layers which abound in non-negligible densities of seeds in the soil (Table 4). Nonetheless, the values of the average densities seeds at the level of the soil layers of the two plots using the Student’s T test revealed no significant difference (p = 0.514) (Table 4).

The values of Shannon index were revealed higher and vary between 2.53 and 1.82, respectively in P08 and P21, contrary to the values of the Pielou equitability index of the two plots. Regarding the soil layers of the two plots, the Shannon index varies from 0.60 to 1.05 in P08 and from 0.42 to 0.88 in P21. However, the Pielou equitability index displays 0.72 to 1 in P08 and from 0.53 to 0.81 in P21. It appears that the soil seed bank of the two plots is less diversified. Consequently, the analysis of the variations of the indices of diversity between soil layers by the nonparametric test of Kruskall Wallis did not revealed any significant difference (p-value > 0.05) between these two indices in the plots studied (Figure 4).

Average numbers of germinated seeds per soil layer demonstrated that soil depth had a significant effect on the germination of the soil seed bank in one of the plots (P08). Indeed, in the plot exploited in 2008, the 0 - 5 cm soil layer stood out from other soil layers by a relatively high germination average per sample (p-value = 0.000). This difference could be explained by the fact that the first centimeters of soil abound more seeds, because they directly receive the seeds from the litter. Our results are similar to those obtained by Daïnou et al. (2011) ; Zebaze et al. (2021) in Cameroon, Douh et al. (2018a) in Congo, Adjalla et al. (2022) in Benin, who have worked in tropical forests. This observation confirms that the soil seed bank would be more abundant in the top 5 centimeters of soil than in the deeper layers (Garwood, 1989; Warr et al., 1993) . While, in the plot exploited in 2021, the differences observed in terms of average germination per sample in each soil layer are not significant (p-value = 0.085).

Which demonstrates that the average germination numbers per sample between different soil layers in this plot are relatively close. Thereby, the 20 - 25 cm layer of soil demonstrated not negligible germination in the two plots studied, hence the interest to consider the latter in the next studies of the soil seed bank.

Globally, the soil seed bank of the two plots started germination from the 2nd week of observation. The average number of germination was higher in the recently disturbed plot (P21) than in that formerly disturbed (P08) with respectively 418 and 347 seedlings germinated during 20 weeks of monitoring. This difference could be explained by the fact that the plot exploited in 2021 was newly disturbed and that the stay of the seeds in the soil bank probably would have started the breaking of dormancy allowing water to reach the seed embryos (Douh et al., 2018b) . In addition, germination peaks were observed during weeks 3 and 9 in each of the two plots.

Our results are different from those obtained by Daïnou et al. (2011) in Cameroon, who observed spontaneous germination of the soil samples taken with a total of 289 individuals germinated in 16 weeks and germination peaks during the 5th, 6th and 7th weeks of the experiment. Whereas Douh (2018) , recorded 61 individuals germinated in 8 weeks and observed germination peaks during the 3 th and 4th weeks. Nonetheless, Zebaze et al. (2021) , after a month of latency, had obtained 230 germinated individuals over a period of 20 weeks, and noticed two peaks of germination during the 12th and 13th weeks. This difference could be explained by the fact that these authors only explored the soil samples in the first 20 centimeters, and that the germination monitoring time has been relatively short compared to the present study. This sighting led us to hypothesize according to which, more recent the disturbance, more the soil seed bank germinates spontaneously in abundance taking advantage of the opening of the canopy.

The density and floristic composition of the soil seed bank in tropical forests are variable.

Hall & Swaine (1980) in Ghana reported that in tropical forests the density of the soil seed bank varies from 100 to 700 seeds/m². Otherwise, in the Amazon rainforest of Brazil, Sousa et al. (2017) quantified 662 seeds/m² in the top three centimeters of soil. In the present study, the densities vary between 1446 and 1742 seeds/m² over the 25 cm soil depth. These results confirm the works of Garwood (1989) , who pointed out that in tropical forests, the abundance of seeds varies from 25 to 3350 seeds/m².

Our results differ from those obtained by other authors. Indeed, Martins & Engel (2007) in Brazil, obtained a density of 800.3 seeds/m² in the first five centimeters of soil. However, in Cameroon, Daïnou et al. (2011) demonstrated densities of 88 to 116 seeds/m² in the top five centimeters of soil. Recently, Zebaze et al. (2021) demonstrate densities that vary between 1099 and 5495 seeds/m² in the first 20 centimeters of the forests of the Dja reserve in Cameroon. While in the Republic of Congo, the works of Douh et al. (2018a) revealed densities of 247 to 330 seeds/m², respectively in “Manilkara forest” and “Celtis forest”.

These two previous studies concerned the first 20 centimeters of soil. These considerable variations of densities obtained could be due to the diversity of methods of characterizing of the soil seed bank, in particular the size of the soil samples collected, the samples collection season, the treatment of the samples and especially the depth of the soil sampled (Garwood, 1989; Warr et al., 1993; Adjalla et al., 2022) .

Nonetheless, Garwood (1989) reminds that the density of the seed bank decreases gradually with the depth of soil. In the present study, this hypothesis concord relatively well in the plot exploited in 2008 (P08) up to 20 cm from of the soil. But, interesting fact, the 20 - 25 cm soil layer demonstrated a superior density to 10 - 15 cm and 15 - 20 cm layers. This a priori surprising observation, suggests that the density of the soil seed bank can increase and/or decrease with the depth of the soil, whatever the biotope considered. On the other side, in the recently exploited plot, it is the 0 - 5 cm layer which displayed a relatively high density, followed of the 20 - 25 cm and 10 - 15 cm layers, while the 15 - 20 and 5 - 10 cm layers demonstrated relatively lower densities. Whence the necessity to explore the 20 - 25 cm soil layer, because it abounds considerable viable seeds potential.

Furthermore, the results of the present study revealed that the plot exploited in 2021 presented a higher percentage of herbaceous species (93.78%) than that exploited in 2008 (82.71%). This could be explained by the fact that the recently exploited forest just got disturbed and, it is the herbaceous species that naturally initiate the healing of after a recent disturbance (Hallé et al., 1978; Daïnou et al., 2011; Vargas et al., 2017) . This could justify the fact that two herbaceous species (Erigeron sumatrensis, IndVal = 0.689 %; p-value = 0.041 and Oldenlandia corymbosa, Indval = 0.618 %; p-value = 0.032) are identified indicator species in this plot.

Otherwise, this predominance of the herbaceous species could be explained by the fact that 1) these plants produce a large quantity of small seeds susceptive to escape at the predation of seed-eating animals (Janzen, 1988) ; 2) these seeds progressively accumulate in large quantities in the soil bank (Borges & Engel, 1993; Arnolds et al., 2015; Vargas et al., 2017) ; and 3) they are able to retain their viability for several decades in the soil bank (Hopkins & Graham, 1987) . These results confirm those obtained in most studies and testify a dominance of the proportions of herbaceous species in the vegetation expressed, i.e. 39.5% in Cameroon (Daïnou et al., 2011) , 41% to 45.3% in Congo (Douh et al., 2018a) .

Two pioneer species of first stage of the plant succession were recorded in the plot exploited in 2008. These include: Musanga cecropioides (2.31%) et Macaranga barteri (0.29%). While, in the plot exploited in 2021, we identified M. cecropioides and Macaranga spinosa to the identical proportions (0.96%). These species have been observed up to 20 cm from the soil in each plot and were more represented in the formerly exploited plot (P08). What suggests that this last would tend towards the process of maturity. But, these low proportions are detrimental to these exploited forests because, according to Hallé et al. (1978) , Hopkins & Graham (1987) , good production of the litter from pioneer species of first stage of the plant succession would be propitious to the installation of long-lived pioneer species such as Nauclea diderrichii, Entandrophragma utile, Triplochyton scleroxylon, Autranella congolensis, many prized by loggers in northern Congo.

Finally, the particularity of the present study resides in the 20 - 25 cm layer which displayed significant proportions of woody species (9%) in the plot exploited in 2008, these are Rubiaceae sp.4 and N. diderrichii. While that exploited in 2021, abounded in 19% of woody species, namely Rubiaceae sp.4, Rubiaceae sp.5 and N. diderrichii, greatly exceeding the proportions obtained in the 15 - 20 cm layer of each of the two plots. The presence of N. diderrichii in the 20 - 25 cm layer of each plot could be explained by the fact that the species produces a large quantity of dormant seeds and small sizes able to retain their viability in the soil bank. In addition, the presence of seeds up to this depth of soil could also be explained by the activity of the soil organisms would involve the seeds into the deep layers of soil (Beaune et al., 2013; Evrard et al., 2017) . Otherwise, Plue & Hermy (2012) assign the presence of seeds in the deep layers of soil to the cracks observed at soil level, to the subsidence of soil around the roots of trees and to rainwater which could convey the seeds in their path of infiltration into the soil.