Metrics details
selective logging removes large trees that are often the main contributors to pollination
We studied pollination patterns of the African mahogany
We investigated two plots in Cameroon corresponding to three tree densities: unlogged forest (Ndama 2002)
a mildly logged forest 1 year after logging (Ndama 2003) and a severely logged forest 30 years after logging (Dimako)
We used four microsatellite markers to perform paternity analysis
Selfing remained below 2% in all treatments
Pollen flow was mainly long distance but with some proximity effects
Average observed within-plot pollination distances were 338
and pollination by trees outside the plots was 70% (Ndama 2002)
Despite sampling a limited number of seeds from a limited number of mother trees
we obtained seeds sired by 35.6–38.3% of the potential within-plot pollen donors
While trees 20 cm in diameter contributed to pollination
results in Dimako suggest that individual larger trees contribute more to pollination than small ones
This effect was not detected in the other treatments
The results suggest extensive pollen flow in Sapelli
the main limiting factor for regeneration after logging may be a reduction in the number of trees capable of producing seeds rather genetic effects due to limits to pollen dispersal
Predicting the fate of genetic diversity of a timber tree species depends on our understanding of the consequences of logging on its reproductive biology
Selective logging of large trees is a common form of harvesting in tropical forests; only trees large enough to reach a minimum diameter cutting limit (MDCL) are cut
forest cover is globally maintained and timber is extracted along roads opened during harvesting operations and then abandoned
the large trees harvested are also likely to have been the main contributors to reproduction
both as seed producers and as pollen dispersers
reducing reproductive tree density through selective logging may have drastic effects on gene flow
and may increase selfing and correlated matings
there is probably always a critical density of flowering trees in a forest below which pollination patterns are severely affected
In view of this uncertainty about the response to natural or man-induced variation in density
predictions about the consequences of selective logging on the reproductive biology of tropical trees are hazardous
If timber harvesting reduces the number of pollinators per seed-bearing tree or increases selfing by delaying fertilization by allo-pollen
then the genetic consequences should be easy to monitor
Genetic relatedness within progenies is expected to increase as a consequence of increased selfing and/or decreased number of pollen donors
levels of allelic and genetic diversity are predicted to show a stronger reduction between fruiting trees and their progenies after logging than before
Analyses of paternity should also provide strong insights into the processes involved
We explored the effect of density reduction due to logging on the reproductive biology of a tropical timber species
Entandrophragma cylindricum Sprague (Meliaceae)
commonly known by its trade name ‘Sapelli’
(1) Does mating occur preferentially between neighboring trees in Sapelli
(2) Does moderate logging limit the number of potential fathers favored by proximity in such a way that seeds from a tree have reduced numbers of fathers
(3) Does severe logging reduce the number trees that are close enough to get a proximity advantage and hence increase the diversity of male parents of the seed crop on a tree
(4) Does low tree density result in increased selfing
we studied Sapelli in two plots in Cameroon
corresponding to three situations (which we henceforth term treatments): one plot before logging; the same plot 1 year after logging – to explore the impact of a limited reduction of density due to selective logging; and a low density plot logged several decades ago – to explore the long-term impact of low density
we estimated the contribution of each mature tree to the fertilization of the seeds produced by focal trees by conducting an analysis of paternity
We also assessed whether mating in Sapelli occurs preferentially between neighboring trees
and whether large diameter trees are the main contributors to pollination
In the continuous tropical forest cover of the Congo Basin
trees are widely spaced and density is typically lower than five flowering individuals per hectare
it is legal to selectively harvest Sapelli trees with a diameter at breast height (DBH) ⩾100 cm
we considered that all trees with a diameter ⩾10 cm DBH were potential pollen donors
as a conservative measure to ensure that all the trees that flowered within the plot had been genotyped
Hence the trees at Dimako are expected to have grown by only 14 cm in DBH since logging
The whole area around Dimako had been logged so we could monitor the effect of large-scale density reduction
Both plots were part of the same continuous forest without any major physical barrier
Densities of small Sapelli trees (10 cm⩽DBH<50 cm) were almost identical in the three treatments
suggesting similar original global densities
Densities of Sapelli trees (DBH⩾10 cm) per hectare in the three treatments were as follows: 1.52 in Ndama before logging (termed ‘Ndama 2002’)
1.13 in Ndama after logging (‘Ndama 2003’) and 0.28 in Dimako
The average diameters of potential pollen donors were 93
Logging intensity was observed to be about 50% of legal-sized trees in Ndama 2002
and was probably much higher in Dimako as both densities and tree diameters were still very low
We examined three subsets of individuals: (1) the whole population of mature trees
termed ‘potential within-plot pollen donors’
from which we extracted (2) a sample of fruiting trees
termed ‘mother trees’ and (3) the progenies of these ‘mother trees’
We sampled seeds under a series of fruiting trees at the end of the fruiting period in Ndama (in 2002 and in 2003)
We selected a subset of focal mother trees
They were chosen to be distributed haphazardly within the plot
our goal was to obtain dispersed distribution of mothers within the plot so as to sample the pollen clouds in different areas
An alternative would have been to choose individuals located at the center of the plots
so as to increase the likelihood of identifying the father of seeds
we would have reduced the independence of the estimates from different mothers
Precision would have been at the expense of generality
we tried to genotype equal numbers of seeds under each female parent
we could analyze the relationship between size of a male tree
its position relative to a focal mother tree and its contribution to the fertilization of seeds
and this independently of female parent crop size
‘male success’ will refer to this estimate
We wanted to highlight neighborhood effects rather that crop size variance effects
Spatial analyses conducted a posteriori (see below for the methodology employed) showed that there were no significant difference in each treatment between the distribution of distances between all pairs of individuals and the distribution of distances between sampled mothers and potential within-plot pollen donors (P=0.089 in Ndama 2002
P=0.52 in Ndama 2003 and P=0.78 in Dimako)
Seeds were collected under crowns of mother trees on the forest floor
All sampled mother trees presented unique combinations of alleles so that we could establish that the ‘mother tree’ was a true parent
Such seeds could either be seeds produced by the maternal tree or were seeds from a neighboring tree sired by pollen from the ‘maternal tree’
Only very few seeds (less than 5%) were not the offspring of the putative mother
This feature demonstrates that most of the times
the putative mother tree was effectively the mother tree
many more cases of genotype discrepancy would have been detected
We retained for further analysis only the mothers for which ⩾ 10 seeds were genotyped without ambiguity
and seeds were then ground to powder in liquid nitrogen
The solution (24 μl total volume) contained 0.4 μl of the four dNTPs (100 μ M)
4.8 μl of 10 × reaction buffer (750 mM Tris–HCl
5.5 μl of mix of the four primer couples (10 μ M)
0.4 μl (2 units) of Taq DNA polymerase (Eurogentec Red GoldStar
One primer from each pair was dye labeled; colors of labels were chosen according to marker size range and did not overlap
Cycling conditions were: denaturing initialized at 94°C for 4 min
30 s at the annealing temperature of 55°C and 1 min at 72°C
with a final extension step of 10 min at 72°C
PCRs were conducted using a PTC100 thermocycler (MJ Research
A total of 2 μl of PCR products were pooled in 15 μl of deionized formamide and 0.2 μl of GeneScan-500XLROX size standard and analyzed on an ABI Prism 310 Genetic Analyser (Perkin Elmer
Size of each fragment was determined using the GeneScan program (Applied Biosystems
Analyses led to three situations: no father identified inside the plot
a single father inside the plot or two or more fathers within the plot among which we were not able to discriminate
We first estimated the percentage of pollination events originating from outside the plot (GFO
gene flow from outside) for each mother sampled
and averaged those values to obtain an estimate for the entire plot
GFO is given as the percentage of each mother's progeny for which no pollen donor was identified inside the plot
We used Spearman's nonparametric rank correlation test to detect whether GFO decreased with distance of the mother tree from the closest edge of the plot (one-tailed test)
We then discarded the few seeds for which two or more potential fathers were found and used only seeds for which a single father was identified (called a ‘known within-plot pollen donor’)
this category comprised more than 90% of the total number of seeds for which at least one potential father was identified within the plot
When a single flowering tree was identified as the father of several seeds from a given mother
we considered this to represent several independent mating events
We computed the percentage of known within-plot pollen donors among the potential within-plot pollen donors
the average DBH of known within-plot pollen donors and the distributions of DBH
We estimated the observed within-plot male mating success (MMSj) of each known pollen donor j as follows:
where ni is the total number of mother trees
the observed number of seeds of mother i fertilized by pollen donor j
that is the average number of seeds sampled under mother trees in that treatment; and SEi the sampling effort for mother tree i
that is the number of seeds sampled from this particular mother
We averaged the MMSj to obtain the MMS at the plot level
Distribution of MMS values in Ndama 2002 was compared to the distributions in Ndama 2003 and Dimako using Kolmogorov–Smirnov tests
To detect whether larger trees were more likely to contribute to pollination than smaller trees
the DBH distributions of effective and potential pollen donors using Kolmogorov–Smirnov tests
To test whether tree density and size distribution affected which size classes contributed to pollination
the DBH distribution of known pollen donors observed in Ndama 2002 was compared to those of Ndama 2003 and Dimako
among within-plot successful pollen donors
large trees fertilized more seeds than small trees
we compared the distribution of MMS values among the five DBH size class: DBH<50 cm
we established which differences among size classes contributed most to this significance by calculating pair-wise Mann–Whitney tests
we compared MMS values among treatment for each DBH class using a Kruskal–Wallis test and when significant we applied Mann–Whitney tests
We tested for directionality of pollen flow using Kolmogorov–Smirnov tests
we determined the angle (relative to north) from each effective and each potential within-plot pollen donor to that tree
We then divided angles into classes (15 degarc) and compared distributions of angles between known within-plot pollen donors and mother trees
and between potential within-plot pollen donors and mother trees
We used Kolmogorov–Smirnov tests to compare the distribution of pollination distances between Ndama 2002 and Ndama 2003
As the plot in Dimako was four times larger than the one at Ndama
we repeated the procedure without taking into account pollination distances above 1397 m
Kinship values were computed in each treatment for each progeny family using as allelic frequencies of the reference population those issued from the whole progeny sample
Kinship values were averaged in each treatment among progeny families to obtain a mean estimate at the plot level
We used Kolmogorov–Smirnov tests to compare the distributions of the kinship values among progeny families between Ndama 2002 and Ndama 2003
We tested differences between He and Fis values among the different subsets using Fstat
Symbols show tree location (triangles=potential pollen donors
DBH<50 cm (this last class of trees were only plotted on the Dimako map); empty triangles
potential pollen donors not identified as effective ones; filled triangles or filled circles (for mothers)
identified effective pollen donor (known pollen donor); empty circles
mothers not identified as pollen donors; arrows
Individual GFO rate is given for each mother close to its location
The estimated percentage of self-pollination was lower than 2% in all three treatments (Table 1) and did not differ significantly among treatments
In Ndama 2003 and Dimako some trees (20 cm in DBH) were recorded as effective pollen donors
Of the potential within-plot pollen donors
38.3 and 36.4% were identified as known pollen donors in Ndama 2002
Average numbers of seeds sired by individual within-plot known pollen donors were 1.47
Individual MMS values tended to increase with decreasing density values from Ndama 2002 to Ndama 2003 and to Dimako
but were only significantly different between Ndama 2002 and Dimako (P<0.001)
MMS and DBH distributions of potential and known pollen donors
DBH distributions of potential pollen donors (white bar) plotted with the DBH distributions of known pollen donors (gray bar)
median values of the MMS of effective pollen donors (MMS) per diameter class (see text for computing details)
Note that the density of small trees (DBH<50 cm) is equivalent in the three treatments
but the numbers are different because the Dimako plot is 4.4 times larger than the Ndama plot
The Kruskal–Wallis tests within treatment showed that MMS values per size class differed only within the Dimako treatment; MMS values in DBH size class (100–119) were higher than in all smaller size classes (class <50
P=0.036) and MMS values in size class (50–79) were higher than in DBH size class <50 cm (P=0.025)
MMS within size class were heterogeneous among treatments for size classes (100–119) and >120 cm (Kruskal–Wallis respectively P=0.002 and P=0.046)
MMS were higher in Dimako than in Ndama 2002 for DBH size classes (100–119) (P=0.006)
and ⩾120 cm (P=0.036) and higher in Dimako than in Ndama 2003 for DBH size class (100–119) (P=0.006)
We detected no significant directionality in pollen flow in any of the three treatments (P=0.091
Distributions of distances between all individuals and between sampled mothers and potential and known pollen donors (pollination distances) ⩾
Distributions of pair-wise distances between all individuals (a-
distributions of pair-wise distances between sampled mothers and potential pollen donors (b-
h-) and distributions of pair-wise distances between sampled mothers and known pollen donors
Class size (under each boxplot) and values of the means (in bold) are reported
The distributions of observed within-plot pollination distances differed significantly between Ndama 2002 and Ndama 2003 (P=0.037) but not between Ndama 2002 and Dimako
The multilocus estimates of outcrossing rates (tm) were 0.98 in the three treatments, while single-locus estimates of outcrossing rate (ts) were 0.96 and were significantly different from the multilocus estimates (Table 1)
estimates of biparental inbreeding (difference tm–ts) were low in the three treatments (tm–ts⩽0.025) but consistent and significant
Distributions of kinship coefficient values among progeny families were not significantly different between Ndama 2002 and Ndama 2003
Genetic diversity parameters for the three treatments are reported in Table 2
We only found a slight but significant difference between values of the standardized allelic richness (A) between within-plot potential pollen donors and progenies in Dimako (P<0.05)
but not between known within-plot pollen donors and progenies in any of the three treatments
The estimates of gene diversity among the different subsets were not significantly different within each treatment
He was high in the three treatments and in all the subsets (He>0.9) while the lowest values of He were found in the progeny subsets in each treatment (He=0.917
Significant but low deficits in heterozygotes were found in all progeny subsets (Ndama 2002
and in the potential pollen donors subset of Dimako (Fis=0.020
The mother trees in Ndama 2003 presented an excess of heterozygosity (Fis=−0.062
paternity analysis demonstrated that trees 20 cm in DBH did contribute to pollination
Hence genotyping ambiguities cannot explain our high GFO values: relatively long-distance gene flow is frequent in Sapelli
There is not a very strong mating advantage to neighboring plants
we observed long-distance pollination within plots
Hence extensive pollen flow is frequent in tropical trees
about one-third of the potential within-plot pollen donors were identified as being effective (35.6
This is a high proportion when taking into account numbers of seeds analyzed and numbers of potential parents within the plots
although MMS has generally been found to be high in most species
each recognized male parent sired on average only 1.47
2.33 and 2.66 of the genotyped seeds in Nadama 2002
Increasing sample size would probably have substantially increased the number of effective parents
This suggests that many more than one-third of adult trees produce male flowers every year; hence
male flowering differs from female flowering and fruit set
analyzing two sets of large Sapelli trees (DBH⩾50 cm
n≈50) in the same locations (Ndama and Dimako) in 1998 and 1999
only one-third of all trees set seeds (unpublished data)
As expected from the numerous observed within-plot pollen donors
kinship within progeny families did not differ among treatments and values of both allelic (A) and genetic (He) diversity in progenies did not differ from those observed in within-plot pollen donors and mother trees
We conclude from these results that despite the flowering-tree density reduction caused by logging
genetic diversity in these populations of Sapelli is not endangered
The extensive pollen flow has sustained the genetic variation
we may assume that even in the low-density treatment
limiting pollen would probably have resulted in increased kinship within progeny families as a result of pollination by a restricted pool of pollen donors
Despite the observed long-distance gene flow
Closely nearby trees were more likely to have pollinated a mother tree
While median within-plot detected pollination distances were 260
the median distance to potential pollen donors were 493
Further GFO decreased or tended to decrease with distance from the edge of the plot
as the GFO rate observed in each treatment showed that more than half of the known pollen donors were situated outside the plot (66–74%)
true average pollination distances in Sapelli are much larger
The significant reduction of observed within-plot pollination distances between Ndama 2002 and Ndama 2003
and the lack of difference between Ndama 2002 and Dimako could be explained by a two-step process in which close neighbors are more favored at intermediate densities but when density becomes even lower there are no more close neighbors
pollination by within-plot neighbors never accounted for more than about one-third of pollination
While there was a slight tendency for mother trees to be heterozygotes
we observed a low (<4%) but consistent deficit in heterozygotes in all progenies
This deficit indicates minor but significant biparental inbreeding and correlated mating
which could be explained by the slight spatial structuring
Indeed limited seed dispersal (almost all seeds collected close to a mother tree effectively came from that tree) associated with some pollination by nearby trees may be sufficient to have generated such a structuring
that this pattern could be explained by selection against even slight inbreeding during the growth of seedlings to seed-producing size
In Ndama, we did not detect any effect of tree size on male reproductive success, but there was a striking one in Dimako. In Dimako, the size distribution of trees was biased toward small trees (Figure 2c)
This resulted in a stronger contribution of small trees to pollination than in Ndama
the large trees (size class (100–119) cm) in DBH were overrepresented among the known within-plot pollen donors and each known within-plot pollen donor of that size contributed to the fertilization of more seeds
This suggests that (1) large trees are more efficient pollen donors than smaller ones
competition among large trees reduces their average success and (3) numbers of seeds sampled and number of small trees were not sufficient to detect whether small trees are less efficient pollen donors
Nevertheless the result depends on a single treatment
there are not enough representatives of each size class to test for whether results could alternatively be explained by the spatial disposition of the trees
We conclude that pollination of Sapelli is a mix of local and long-distance pollen flow
Long-distance pollen is sufficiently prevalent to have a strong homogenizing effect on the regional gene pool in the diffuse Sapelli populations
whatever the local densities of flowering individuals
reduced densities after logging resulted in a dramatic increase in selfing
predicting the consequences of logging policies and regulations requires genetic studies on the species in question
La flore forestière de la Cote d’Ivoire Publication no
Plant pollinator interaction in tropical rain forests
genetic differentiation and speciation in tropical rain forest plants
Controlling the false discovery rate – a practical and powerful approach to multiple testing
Microsatellite primer amplification by multiplexing: a first application to Eucalyptus grandis
Effects of habitat fragmentation on the reproductive ecology of four plant species in mallee woodland
Simple sequence repeats provide a direct estimate of pollen-mediated gene dispersal in the tropical tree Gliricidia sepium
Population structure delineated with microsatellite markers in fragmented populations of a tropical tree Carapa guianensis (Meliaceae)
L’analyse des cernes : applications aux études de croissance de quelques essences en peuplements naturels de forêt dense africaine;
Mating system of Carapa procera (Meliaceae) in the French Guiana tropical forest
Diamètre de fructification de quelques essences en forêt naturelle centrafricaine
Evidence of low gene flow in a neotropical clustered tree species in two rainforest stands of French Guiana
High level of genetic differentiation for allelic richness among populations of the argan tree (Agrania spinosa (L) Skeels) endemic to Morocco
Genotyping of mature trees of Entandrophragma cylindricum with microsatellites
FaMoz: a software for parentage analysis using dominant codominant and uniparentally inherited markers
Comparison of microsatellites and amplified fragment length polymorphism markers for parentage analysis
Disturbance-induced density-dependent seed set in Shorea siamensis (Dipterocarpaceae) a tropical forest tree
FSTAT (Version 12): a computer program to calculate F-statistics
Classification and ecology of closed-canopy forest in Ghana
Effect of forest fragmentation on genetic diversity and mating system in a tropical tree Pithecellobium elegans
The breeding structure of tropical tree populations
SPAGEDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels
Ecological profiles of Ghanaian forest trees
Viability selection at three early life stages of the tropical tree
Pollen-mediated gene flow and differential male reproductive success in a tropical pioneer tree Cecropia obtusifolia Bertol (Moraceae): a paternity analysis
Variation in pollen dispersal between years with different pollination conditions in a tropical emergent tree
Estimation of outcrossing rate on Dryobalanops aromatica Gaernt F in primary and secondary forests in Brunei Borneo Southeast Asia
Long-distance pollen flow and tolerance to selfing in a neotropical tree species
Mating system parameters of Dryobalanops aromatica Gaertn f (Dipterocarpaceae) in three different forest types and a seed orchard
Mating system parameters in a tropical tree species Shorea leprosula Miq (Dipterocarpaceae) from Malaysian lowland dipterocarp forest
Spatial genetic structure of a tropical understory shrub
Isozyme variation in tropical trees: patterns of genetic organization
Statistical confidence for likelihood-based paternity inference in natural populations
The effect of the density of flowering individuals on the mating systems of nine tropical tree species
Microsatellite analysis of the breeding system and seed dispersal in Shorea leprosula (Dipterocarpaceae)
Reproductive and genetic consequences of forest fragmentation: two case studies of neotropical canopy trees
Genetic diversity and outcrossing rate between undisturbed and selectively logged forests of Shorea curtisii (Dipterocarpaceae) using microsatellite DNA analysis
GENEPOP (version 12): population genetics software for exact tests and ecumenicism
Extensions of models for the estimation of mating systems using n independent loci
Pollen dispersal in low-density populations of three neotropical tree species
The flower biology of the Meliaceae and its bearing on tree breeding
Cytology and reproductive biology of Meliaceae
A review of outcrossing and pollen-mediated gene flow in neotropical trees
Increased pollen flow counteracts fragmentation in a tropical dry forest: an example from Swietenia humilis Zuccarini
Download references
We are grateful to A Mbenda and the field workers for help in collecting material in the field
M-P Dubois and C Debain for help in the laboratory work
S Gerber for helpful comments on the paternity analysis
P Jarne and D McKey for comments on the manuscript
This work was supported by funds from Cirad (Centre de Coopération Internationale en Recherche Agronomique pour le Développement)
This paper is dedicated to late Caroline Dubois who started the field observations on the sites
This paper is part of the PhD of Mathieu Lourmas at the University of Montpellier
on the impact of human activities on population genetics and conservation biology of Entandrophragma cylindricum in the Congo Basin
The authors are participating in several projects that aim at understanding genetic population processes in tropical trees and apply these results to their conservation
Download citation
DOI: https://doi.org/10.1038/sj.hdy.6800976
Anyone you share the following link with will be able to read this content:
a shareable link is not currently available for this article
the charity association of the First Lady also offered diverse gifts to the population in the locality on Saturday
Наслаждайтесь азартом на любом устройстве! Скачайте мобильное приложение Vavada kz и играйте без блокировок
Начните с приветственного бонуса и открывайте для себя турниры
Offer your company and your employees the best information
Take advantage of our preferential rates reserved for professionals
Discover our digital subscription offers to find Cameroon-Tribune at home
Statistics from Doume Health District in the East Region show that greater awareness and the offer of motorbike ambulances by UNFPA have enabled more rural women to be promptly transported to health facilities for safe delivery
According to the last Cameroon Demographic and Health Survey and the 2014 Multi-indicator Cluster Survey Statistics
about 7,000 women die in the country due to pregnancy complications or during delivery
While neonatal mortality (the death rate of babies in the first 28 days of life) was 28 deaths per 1,000 live births
with the support of the United Nations Children’s Fund
UNICEF and the United Nations Population Fund
the project is funded by the Islamic Development Bank
It covers 34 health districts in the Far North
East and South Regions - the most vulnerable in the country
which comprises three administrative subdivisions in Upper Nyong Division (Doume
The health district counts 132 communities and a population of 55,000
Two years into the implementation of the project in Doume Health District
says a 100 per cent increase was recently recorded in vaccination and antenatal clinic attendance
This followed sensitisation campaigns for people to continue to patronise orthodox health services – despite the Coronavirus pandemic
The awareness campaigns were led by Multitask Community Health Workers and volunteers of the Ministry of Youth Affairs and Civic Education
the number of births in health facilities in Doume Health District shot up from 25 per cent two years ago to 55 per cent
and this happened at home where she gave birth
there were three newborn deaths – two in hospital and one at home
The ones who died in hospital were brought in late
we recorded between 15-28 declared deaths of children and 5-6 maternal deaths a year,” the District Medical Officer notes with a feeling of “work well done!”