A general, community-level test of the Janzen-Connell hypothesis C. E. Timothy Paine 1 , Natalia Norden 2 , Christopher Baraloto 3 , Jérôme Chave 4 , Pierre-Michel Forget 5 1 ! IEU, Universität Zürich, Switzerland 2 ! Departamento de Ciencias Biológicas, Universidad de los Andes, Colombia 3 ! INRA, UMR "Ecologie des Forêts de Guyane", French Guiana 4 ! Université Paul Sabatier, CNRS, France 5 ! Muséum National d’Histoire Naturelle, France
Crax alector eating seed of Lecythis persistens
Janzen-Connell (JC): Increased diversity by increasing seedling-adult distances Seed Seed/seedling rain survival Seedling Recruitment A form of negative density dependent performance, mediated by (relatively) host-specific natural enemies
Lots of evidence for JC Seedling survival is reduced... ! Near conspecific adults ! In crowds of conspecific seedlings ! ...Often because of host-specific insects or pathogens Seedling escape from fungalpathogens 28r Tree I * . 80 - . . . - . :. . .. . . . . - 60 + * .... 0 4 * 0 . . . . . - . c 0 . 5 20 . . Tree I . .. -0 h 0 - Â m c A 0 1 LTree 3l 4 Tree 4 7 201 : . : * Tree 3 Distance south of parent tree (m) Distance south of parent tree (m) Augspurger, 1983, J. Ecology FIG. 3. Percentage of seedlings of Platypodium elegans on Barro Colorado Island, Panama, FIG. 1. Seedling density in relation to distance south of four parent trees of Pla[vpodium elegans dying from damping-off during the first 3 months as a function of distance along transects south on Barro Colorado Island, Panama. of four parent trees. The lines show the percentage damping-off predicted by the regression if seedling density and light environment at each distance were at the average value for the transect. Points are the observed levels of damping-off, corrected for variation due to seedling density and light environment, i.e. 1. Differences among the four trees of Platjyodium elegans on Barro TABLE + b3('Â¥Lsa -x~(sa.p))> = yi(obs) "~(corr) + bz('denslty - X ~ d e n s i t y ) ) are the partial regression coefficients for X , + X , respectively, from the multiple where 6, and b3 Colorado Island, Panama, in distance and density of seedlings south of the regression presented in Table 2). parent, and incidence of death of seedlings by damping-off during first 3 months after germination. 3 Tree 1 2 4 Number of seedlings observed 1884 926 298 115 Median distance of seedlings 22 I3 16 12 I Tree I Tree 2 from parent (m) Seedling density at median 25.0 12.0 2.7 1.5 distance (m ') 83.9 46.1 Seedlings dying (%I) 85.2 80.0 34.8 Seedlings dying due to 81.2 63.0 54.0 damping-off (96) Proportion of seedling deaths due 95.2 78.7 64.4 75.5 to damping-off ('XI) first 3 months was higher within 20 m of the pankt than at any greater distance (Fig. 2). Distance south of parent t r e e ( rn For seedlings of Trees 1 and 2, mortality was higher at intermediate than at far distances FIG. 4. Percentage of seedlings of Platypodium elegans on Barro Colorado Island, Panama, (Fig. 2). For Tree 1, seedling mortality was lowest in the large gap at 35-55 m; for Tree 2, dying from damping-off during the first 3 months at selected distances from Trees 1 and 2. An mortality at far distances was slightly lower than in the small gap at 5-10 m (Fig. 2). The equal number of seedlings were observed at each distance interval. Shaded bars represent gaps. The initial mean density of seedlings (m-l) occurs above each bar. four trees differed in the proportion of seedlings surviving the first 3 months (Table 1); mortality increased with seedling density. Mortalify factop Causes of seedling mortality were determined only in the first 3 months when damping-off was the predominant cause. It accounted for 64% (Tree 3) to 95% (Tree 1) of seedling deaths (Table 1). The remaining mortality was unevenly divided among six other factors: (i) the fall of branches and leaves onto the seedling; (ii) poor root establishment;
Why return to JC when so much has already been done? Most studies performed on few species Studies of recruitment beneath con-/hetero- specifics are limited to species common enough to provide decent sample size What can we add? ! A community-wide, taxonomy-free analysis ! Using functional traits and evolutionary history
Three ways natural enemies could find seedlings Possibility Caveat Strictly host-specific Unlikely Similar evolutionary Novel adaptations may history make closely-related species ecologically divergent Similar functional traits Which are relevant to measure? Also, environmental filtering shapes trait distributions
Objective: test generality of JC across entire tree community Why? Lots of evidence for JC for single species, ! generality is little known How? Functional and phylogenetic similarity ! of seedlings and adults Who? Fabulous collaborators! Pierre-Michel Forget: Seeds Chris Baraloto: Traits Natalia Norden: Seedlings Jérôme Chave: Phylogeny
Sampling of adults (3626) and seedlings (5247) 98 seed collectors, with 202 seedling plots 0$'+1234#*%+% !"#$%&#'( , % -$%.*/ + * $ # )
Sampling functional traits: 7 traits on each of 4672 trees (668 spp) Leaf traits: &'& "#$ ! Chlorophyll content ('& ()*+,*-)* ! Leaf toughness ./0)12 &(D ! SLA ? $#% 56780)12 > ; ! C:N "< = $#$ ()*+ Stem traits (50 ()*+A): ! Sapwood density =; 3*-4 =@AB !$#% ./92: 56783*-4 5BC8@ ! Trunk moisture ! Bark thickness !"#$ For Review Only Baraloto et al, in press, Ecol Lett
Community phylogeny 611 species of French Guiana rain-forest trees using matK + rbcL Gonzalez et al 2010 PLoS ONE , J. Ecology
Modeling seedling survival Cox proportional hazard log(seedling size) ( ) Seedling ~ + log(seedling density) survival + log(adult basal area) * dissimilarity Stem traits { } Dissimilarity = Leaf traits Phylogeny Also stratified by recruitment year
Bigger seedlings survive better Seedling density has little effect Median expected survival (years) 4 4 3 3 2 2 1 1 Ht: <0.0001 N sdl: 0.0211 0 0 1 5 10 20 50 15 30 50 80 Seedling height (cm) Seedling density (m-1) No important interactions between seedling height and dissimilarity
Adult-seedling similarity affects survival 4 4 3 3 2 2 Median expected survival (years) Basal area BA: ns BA: ns 3.8 D: 0.0109 D: ns 1 5.4 1 BAxD: 0.0063 BAxD: ns 6.1 6.6 8.1 0 0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 2 3 4 5 Leaf trait dissimilarity ALL trait dissimilarity 4 4 3 3 2 2 BA: 0.0023 BA: 0.0096 D: ns D: 0.0025 1 1 BAxD: 0.0152 BAxD: 0.0023 0 0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 180 220 260 300 Stem trait dissimilarity Phylogenetic dissimilarity
How to explain? 4 4 3 3 2 2 Median expected survival (years) Basal area BA: ns BA: ns 3.8 D: 0.0109 D: ns 1 1 5.4 BAxD: 0.0063 BAxD: ns 6.1 6.6 8.1 0 0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 2 3 4 5 Leaf trait dissimilarity ALL trait dissimilarity 4 4 Adults with... 3 3 2 2 BA: 0.0023 BA: 0.0096 D: ns D: 0.0025 Dissimilar leaves ↑ survival ! 1 BAxD: 0.0152 1 BAxD: 0.0023 0 0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 180 220 260 300 Dissimilar stems ↓ survival ! Stem trait dissimilarity Phylogenetic dissimilarity ! (thus all traits together have no effect) Dissimilar evolutionary history ↑ survival !
How to explain? 4 4 3 3 2 2 Median expected survival (years) Basal area BA: ns BA: ns 3.8 D: 0.0109 D: ns 1 1 5.4 BAxD: 0.0063 BAxD: ns 6.1 6.6 8.1 0 0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 2 3 4 5 Leaf trait dissimilarity ALL trait dissimilarity 4 4 JC predicts greater survival 3 3 2 2 BA: 0.0023 BA: 0.0096 D: ns D: 0.0025 where adults very dissimilar 1 BAxD: 0.0152 1 BAxD: 0.0023 0 0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 180 220 260 300 from seedlings Stem trait dissimilarity Phylogenetic dissimilarity Support for JC for leaf traits and phylogeny NOT stem traits Environmental filtering on stem traits may counteract natural-enemy-mediated effects
Conclusion Strong support for Janzen-Connell hypothesis at the community level for leaf traits and phylogeny Confirms a long chain of literature Shows generality of JC hypothesis, even for rare spp Environmental filtering also limits seedling survival
What’s next? Trait/phylogenetic similarity of neighboring seedlings Repeat with minimum (rather than mean) dissimilarities or distance-weighting of adults Simultaneous evaluation of environmental filtering and JC?
Recommend
More recommend
