Are we missing the forest for the trees? Conspecific negative density dependence in a temperate deciduous forest

Kathryn E Barry; Stefan A Schnitzer

doi:10.1371/journal.pone.0245639

. 2021 Jul 15;16(7):e0245639. doi: 10.1371/journal.pone.0245639

Are we missing the forest for the trees? Conspecific negative density dependence in a temperate deciduous forest

Kathryn E Barry ^1,^*, Stefan A Schnitzer ²

Editor: RunGuo Zang³

PMCID: PMC8282035 PMID: 34264937

Abstract

One of the central goals of ecology is to determine the mechanisms that enable coexistence among species. Evidence is accruing that conspecific negative density dependence (CNDD), the process by which plant seedlings are unable to survive in the area surrounding adults of their same species, is a major contributor to tree species coexistence. However, for CNDD to maintain community-level diversity, three conditions must be met. First, CNDD must maintain diversity for the majority of the woody plant community (rather than merely specific groups). Second, the pattern of repelled recruitment must increase in with plant size. Third, CNDD should extend to the majority of plant life history strategies. These three conditions are rarely tested simultaneously. In this study, we simultaneously test all three conditions in a woody plant community in a North American temperate forest. We examined whether understory and canopy woody species across height categories and dispersal syndromes were overdispersed–a spatial pattern indicative of CNDD–using spatial point pattern analysis across life history stages and strategies. We found that there was a strong signal of overdispersal at the community level. Across the whole community, larger individuals were more overdispersed than smaller individuals. The overdispersion of large individuals, however, was driven by canopy trees. By contrast, understory woody species were not overdispersed as adults. This finding indicates that the focus on trees for the vast majority of CNDD studies may have biased the perception of the prevalence of CNDD as a dominant mechanism that maintains community-level diversity when, according to our data, CNDD may be restricted largely to trees.

Introduction

Conspecific negative density dependence (CNDD) is one of the most empirically supported mechanisms for the maintenance of plant species diversity [1–4]. Conspecific negative density dependence occurs when small individuals have relatively low rates of growth and survival near adult members of their own species (conspecifics). This constraint on growth near conspecifics results in a distinct spatial pattern where adult conspecifics occur further away from each other than would be expected by chance (overdispersion; [5]; but see [6,7]). Thus, CNDD is predicted to result in stable species coexistence across the landscape because dominant species cannot displace subordinate ones [8,9]. Evidence for CNDD has been reported in a variety of ecosystems, including lakes, deserts, grasslands, marine ecosystems, and particularly in temperate and tropical forests [1,2,7,10–16]. For forests, over the past ten years alone, evidence for CNDD has been reported more than 30 times in 13 countries across five continents [17].

However, strong evidence for CNDD at the seedling level may not result in maintenance of diversity at the forest level if NDD reduces clustering of seedlings but does not overcome the initial clumped pattern of seedlings around adults (due to dispersal limitation) [18]. That is, if CNDD does not result in a pattern of overdispersion, then it may fail to stably maintain species diversity ([5,19,20]; but see [21,22]). For CNDD to be a likely mechanism maintaining community-level diversity in temperate forests, the following three conditions must be met. 1) Individuals of the majority of the species in the community will be overdispersed because of greater mortality near conspecific adults. If only a small proportion of the species are overdispersed, then CNDD may not theoretically benefit rare species enough to maintain diversity [8,9]. 2) The degree of overdispersion will increase with ontogenetic (life-history) stage. That is, the signal of CNDD should compound as individuals mature, and thus larger individuals of any given species should be more overdispersed than smaller individuals (see Zhu et al. 2015 [18]). 3) CNDD will operate across life history strategies, including species that vary in growth form and dispersal syndrome. If CNDD is the main mechanism driving diversity maintenance, as suggested by previous studies (e.g., [3,17]) then it should operate on the plant growth forms that contain the highest diversity. If these three conditions are met, then CNDD is likely to be sufficiently strong to maintain community-level diversity.

Previous studies may have overestimated the importance of CNDD in forest ecosystems for two reasons. First, the vast majority of studies that examined CNDD in vascular plant species focused on growth and mortality at the seedling stage [1,2,14,16,23,24]. Dynamics at the seed-to-seedling and seedling-to-sapling transitions do not necessarily translate to overdispersion in the larger size classes [25] and may overestimate the role of CNDD [26]. Due to dispersal limitation, most seeds arrive beneath the parent tree and thus most seedlings are also congregated there. CNDD can maintain diversity only if seedling mortality beneath the conspecific is sufficiently strong to overcome and reverse the initial clumped seedling distribution. Furthermore, CNDD must be high enough to exceed the null expectation of high seedling mortality near the parent tree purely because there are more seedlings present [27]. If so, the negative effects of growing near a conspecific adult should compound as individuals mature. As individuals grow, they compete more intensely with adults or acquire more pathogens or both, and thus the level of overdispersion should increase with plant size.

Currently, the evidence for CNDD beyond the seed to seedling transition is mixed. For example, a study by Yao et al. [25] found that CNDD decreased with increasing tree ontogeny in a temperate forest. In fact, many species in both temperate and tropical forests do not have an overdispersed distribution [19,28]. By contrast, Guo et al. [29] found that 75% of tree species demonstrated CNDD as adults in subtropical forests (see also [30–32]).

Second, the vast majority of CNDD studies in forests focused only on trees, ignoring other important plant growth forms (e.g., [5]). The selection of trees to test CNDD as a general mechanism for the maintenance of diversity is particularly problematic in temperate forests, where canopy tree species represent a relatively small fraction (~7%) of the total vascular plant community [33]. Furthermore, canopy trees may be more prone to overdispersion due to their capacity for long distance dispersal [34–36] and thus they may bias our understanding of the importance of CNDD for the diversity maintenance of the larger plant community. By contrast, understory plants (including understory woody species) represent a larger share of diversity but have a lower capacity for long distance dispersal due to their relatively short stature and position in forest understory. Furthermore, few understory plant species have dispersal syndromes that favor long distance dispersal [33]. Many understory species are gravity dispersed while the majority of temperate canopy trees are wind dispersed. Thus, the strength of CNDD may interact with plant dispersal syndrome.

Nonetheless, if CNDD is the primary mechanism that maintains community level diversity, we would expect it to operate across life history stage and life history strategy. We addressed these three core conditions for CNDD to be a general mechanism for the maintenance of plant species diversity by evaluating the spatial patterns of a woody plant community across life history strategies (shrubs, understory trees, mid-story trees, canopy trees, and lianas) and ontogenetic stages (seedling, sapling, and adult) in the field in a temperate forest in western Pennsylvania, USA. We tested three specific hypotheses: 1) Woody plant diversity in temperate forests is maintained by CNDD, and thus we predict that the majority of plant species will be overdispersed. 2) The effects of CNDD compound as plants grow, and thus overdispersion will increase with plant size. 3) CNDD operates independently of growth form and dispersal syndrome, and thus we predict that the pattern of overdispersion will be found in the majority of the species of all plant groups. We tested these hypotheses by examining the degree of overdispersion in a woody plant community, which included a range of plant life-history stages (i.e., sizes) and life-history strategies (i.e., growth form and dispersal syndrome).

Materials and methods

Study site

We conducted this study at Powdermill Nature Reserve with permission from the Carnegie Museum of Natural History. Powdermill Nature Reserve is an 890-hectare reserve located in the Allegheny plateau at the base of the Appalachian Mountains in southwestern Pennsylvania, USA (Westmoreland County; 40^o09’ N, 79^o16’ W). This region receives ~1100 mm of precipitation per year and is characterized by mixed mesophytic vegetation that is dominated by maples (Acer spp.), tuliptree (Liriodendron tulipifera), and oaks (Quercus spp.). Elevation at Powdermill Nature Reserve ranges from 392 to 647 m above sea level. Powdermill Nature Reserve contains a matrix of vegetation types consisting primarily of secondary deciduous forest but with several areas of maintained fields and managed lands. Last known logging occurred in this region in the 19^th century, and land was primarily used for agriculture into the early 20^th century (see [37] for more detailed site description).

Plot establishment and plant census

In May and June of 2014, we established sixteen 10-m diameter circular plots in the >90-year-old secondary temperate deciduous forests at Powdermill Nature Reserve. We chose the 10-m diameter spatial grain because this size was thought to be a suitable size to test for spatial patterns associated with NDD in a Malaysian forest ([38]; see also [19,28]). We avoided canopy gaps for the placement of each plot, and each plot had >80% canopy cover. We ensured that the plots were not within 10 m of a waterway, that soil cover was not predominantly rocks, and that plots were at least 20 m from any edge. We used a Trimble GeoExplorer 6000XH to measure the precise location (up to 10 cm accuracy) of all woody plant individuals >10 cm height in each plot (Trimble Navigation Limited, Westminster, CO). For each individual, we measured height and basal diameter, and we identified them to species.

To examine how overdispersion changes with plant size, we divided individuals into four height classes (<0.5 m, 0.5–1 m, 1–5 m, and 5–10 m). We use these height classes as a proxy for both relative age (assuming that plants get taller as they get older) and position within the forest. Individuals that are shorter are less likely to be able to disperse seeds farther away than individuals that are taller even if they belong to the same species and have the same dispersal syndrome. To understand how overdispersion interacts with life-history strategy, we classified each species as either canopy or understory (growth form), and as either bird, wind, self, or other animal dispersed (dispersal mechanism) based on species descriptions in the Flora of North America [39].

Data analysis

We performed all data analysis in R statistical computing software [40]. To measure plant spatial distribution (the degree to which plants are clustered or overdispersed), we calculated Ripley’s K in the package “spatstat” using Ripley’s translational border correction at each plot for each species and then, for ease of interpretation, converted K to Besag’s L [41–44]. Several studies have demonstrated that spatial point pattern analysis is capable of detecting spatial patterns that can be attributed to mechanistic processes (e.g. [5,45]). To eliminate point patterns based on low replication, we removed species at any plot with fewer than five individuals as point patterns with fewer than 5 points in our data had significantly larger variance than those with >5 points. We then calculated a pooled L for each comparison (by species, by growth form, by growth form/plant size, or by growth form/dispersal mechanism) by weighting the individual L estimates by the number of points in a given L-function (methods follow [46]). For the growth form by dispersal mechanism interaction, we limited the final analysis to bird and wind dispersed species because these two groups had sufficient replication for robust comparisons between canopy and understory plants.

We bootstrapped these estimates 999 times to create 95% confidence intervals. We then calculated the predicted L for complete spatial random to compare our spatial patterns to complete spatial random. Data manipulation of input to and output from point pattern analysis was done using a combination of the “abind”, “gridExtra”, and “reshape” packages [47–49]. We constructed all figures in the package “ggplot2” [50].

To more easily interpret the figures, we corrected our measures of L with the distance at which each measure of L was calculated (L(d)-d). Besag’s L is a measure of spatial aggregation, and when L(d)-d is positive, a greater proportion of neighbors are observed within distance d of focal individuals than predicted by a complete spatial random pattern. When L(d)-d is negative, a smaller proportion of neighbors are observed within distance d of focal individuals than predicted by a complete spatial random pattern (Fig 1, [51]). We considered plants to be overdispersed based on their point pattern when the linear regression slope of L(d)-d was significant and positive (using the command “lm” in R, S1 and S2 Tables) with increasing distance (d) [46]. This designation implies that more individuals are found far away from an individual of a given species than near an individual of that species. We considered plants to be clustered based on their point pattern when the linear regression slope of L(d)-d was significant and negative with increasing distance. These designations differ from “pure overdispersion” (i.e. regularity or inhibition), which would begin with a significantly negative L(d)-d that indicates fewer individuals close to the parent than would be expected by chance (Fig 1B and 1C, [44,46,52]). However, natural dispersal typically results in more conspecific seeds and seedlings close to adults than predicted by complete spatial random, and thus we did not expect to find a significant negative L(d)-d of seedlings close to the parent [53]. Therefore, we accounted for dispersal limitation by focusing on overdispersion as having a positive slope with regards to distance (d), which indicates a significant increase in individuals with distance from the adult, at the community scale. However, CNDD should result in increasing overdispersion with plant size as the effects of CNDD compound with time (as the plant matures). Thus, we might expect a spatial signature of “pure overdispersion” in larger size classes if CNDD is capable of overcoming initial dispersal patterns and thus stably maintaining coexistence.

Fig 1 — a.) Conceptual diagram of interpretations of different quadrants of spatial point pattern space for Besag’s L after it has been centered by distance. b.) Hypothetical distribution where individuals are perfectly overdispersed, there are more individuals than expected from complete spatial random but only at larger distances, i.e. those individuals are farther apart. c.) Hypothetical distribution where individuals are perfectly underdispersed. There are more individuals than expected from complete spatial random but only at shorter distances, i.e. those individuals are close together.

We considered any point pattern to be significantly different from complete spatial random if a mixed effect linear model including all factors of the L(d)-d and the distance was significant (calculated using the command “lmer” in package “lme4” with plot as a random effect and using the package “lmerTest” to calculate p-values; Table 1; [54,55]). Further, we include the results of the same models but with a more conservative estimate of degrees of freedom in S2 Table. If the total model was considered significant, we did not consider the point pattern to be significantly different from complete spatial random at any distance where the bootstrapped 95% confidence intervals of L(d)-d overlap with complete spatial random. We considered any two point patterns to be significantly different from each other if their bootstrapped 95% confidence intervals did not overlap at a given distance.

Table 1. Results of mixed effects linear model to calculate L statistic significance at Powdermill Nature Reserve in Southwestern Pennsylvania for all comparisons of all individuals >10 cm height.

Model	dF	T Stat	P value	Figure
All individuals	2228	19.37	<0.0001	1a
All individuals, <0.5	781	-4.477	0.6689	1b
All individuals, 0.5–1	889	7.83	<0.0001	1b
All individuals, 1–5	797	8.5828	<0.0001	1b
All individuals, >5	358	-14.97	<0.0001	1b
Understory	1071	10.191	<0.0001	2a
Canopy	1059	5.796	<0.0001	2a
Canopy, <0.5 m tall	380	-7.14	<0.0001	2b
Canopy, 0.5–1 m tall	478	8.67	<0.0001	2b
Canopy, 1–5 m tall	463	9.14	<0.0001	2b
Canopy, >5 m tall	186	-10.71	<0.0001	2b
Understory, <0.5 m tall	397	2.045	0.0415	2c
Understory, 0.5–1 m tall	407	0.3348	0.7379	2c
Understory, 1–5 m tall	381	1.038	0.2998	2c
All individuals, bird dispersed	1159	2.86	0.0042	3a
All individuals, animal dispersed	200	-6.806	<0.0001	3a
All individuals, wind dispersed	803	3.404	<0.0001	3a
Overstory, wind dispersed	601	3.903	<0.0001	3b
Overstory, bird dispersed	250	5.482	<0.0001	3b
Understory, wind dispersed	199	-8.895	<0.0001	3b
Understory, bird dispersed	906	-3.533	0.0004	3b

Open in a new tab

To calculate significant differences from complete spatial random we used a mixed effects linear model with plot as a random effect to control for between plot differences due to environmental heterogeneity between plots. We report model degrees of freedom based on the number of L estimates (calculated every 10 cm per point pattern per plot). We report any pooled point pattern as overdispersed if it has a significantly positive slope and any pooled point pattern as clustered if it has a significantly negative slope.

Results

At the community level, all woody plants combined were significantly overdispersed (Fig 2A). The largest individuals (>5 m tall) had a significantly lower overdispersion (L(d)-d) at intermediate distances (2-5m), than the two middle height size classes (1m – 5m and 0.5–1 m); however, L(d)–d did not differ significantly among the larger size classes at distances greater than 5 m (Fig 2B). By contrast, the smallest individuals (< 0.5 m) had significantly lower overdispersion than intermediate height individuals (0.5-1m and 1–5 m tall) for all distances greater than 2m (Fig 2B), and significantly lower dispersion than individuals in all of the larger height categories for distances greater than 5m. Thus, all but the smallest size classes were overdispersed at longer distances from the adult tree, indicating that, at the community-level, NDD was strong enough to overcome the initial clumped distribution of seedlings as the plants grew.

Both canopy trees and understory plants were significantly overdispersed; canopy trees were more overdispersed (significantly higher L(d)-d) at distances greater than 3 m (Fig 2A). The differences in overdispersion between canopy and understory plants become more pronounced with plant life history stage (i.e., plant size). Canopy trees did not differ significantly from complete spatial random when they were small and young, but became significantly overdispersed when they were larger (Fig 3B), which is consistent with CNDD. Understory plants displayed the opposite pattern: they were overdispersed when small, but larger individuals were indistinguishable from complete spatial random (Fig 3C).

Fig 3 — Black dotted line throughout represents the complete spatial random prediction. a.) Canopy (n = 29 point patterns) and understory plants (n = 33 point patterns) were both significantly overdispersed, indicative of negative density dependence. Canopy plants were significantly more overdispersed than understory plants. b.) Canopy plants were more overdispersed with increasing life-history stage in accordance with predictions for negative density dependence (n_<0.5 = 14 point patterns, n_0.5–1 = 48 point patterns, n_1-5 = 32 point patterns, n_>5 = 15 point patterns). c.) Understory plants were not more overdispersed with life-history stage (n_<0.5 = 21 point patterns, n_0.5–1 = 22 point patterns, n_1-5 = 20 point patterns). Grey shaded regions represent 95% confidence intervals, darker grey regions represent overlapping confidence intervals. When confidence intervals overlap, we consider two point patterns to be the same in the overlapping region. We consider point patterns where the confidence intervals overlap with the black dotted line to not be significantly different from complete spatial random in that region.

All of the four dispersal mechanisms that we examined, wind, bird, and self-dispersed species were overdispersed and statistically indistinguishable from each other. Species dispersed by animals other than birds (including secondary dispersal by squirrels) were all significantly less overdispersed than the other three dispersal types (Fig 4A). Dispersal syndrome for bird and wind dispersed species did not explain the differences in spatial pattern between canopy trees and understory plants; canopy trees were always more overdispersed than understory plants regardless of dispersal mechanism (Fig 4B), suggesting that the height of canopy trees is the most important factor in dispersal distance.

Fig 4 — a.) Wind dispersed(n = 16), bird dispersed (n = 23), and self dispersed(n = 2) species were significantly more overdispersed than species dispersed by animals other than birds(n_animal = 6). b.) Canopy plants were significantly more overdispersed than understory plants regardless of dispersal mechanism(n_canopy-bird = 5, n_canopy-wind = 12, n_{understory-bird} = 18, n_{understory-wind} = 5). Bird dispersal was emphasized here; however, plants did not differ significantly from wind dispersed plants either in the canopy and the understory. All reported sample sizes (n) are in number of total point patterns contributing to a pooled L function. Grey shaded regions represent 95% confidence intervals, darker grey regions represent overlapping confidence intervals. When confidence intervals overlap, we consider two point patterns to be the same in the overlapping region. We consider point patterns where the confidence intervals overlap with the black dotted line to not be significantly different from complete spatial random in that region.

Discussion

We found that canopy trees were overdispersed and the strength of overdispersion increased with tree size–two critical conditions for CNDD to be a general mechanism for the maintenance of woody plant species diversity. Increasing overdispersion with increasing plant size is predicted by CNDD because plants should survive and grow best away from conspecific adults due to intraspecific competition [56] or the negative effects of natural enemies [8,9,57,58]. Our findings are consistent with a growing number of studies that have reported that CNDD is a viable mechanism to maintain canopy tree diversity in temperate and tropical forests (e.g., [1,2,7,14,32,38]). Thus, our findings support CNDD as a mechanism for the maintenance of canopy tree species diversity. However, we cannot rule out the possibility that trees may be more likely to be overdispersed with size simply because the larger (and presumably older) the tree the greater the probability of mortality for the parent (which is often nearby) [59].

For woody understory plants, our spatial patterns did not meet the criteria for CNDD to maintain species diversity. Understory species were overdispersed only in the smallest size classes, and overdispersion did not increase with plant size, which we use as a proxy for life history stage. If CNDD is operating in understory plants in these forests, it does not appear to be sufficiently strong to overcome the initial clumped dispersal pattern of seedlings, and therefore it did not result in overdispersion. Similar conclusions that CNDD may not be a general mechanism for the maintenance of non-tree plant diversity were reported for tropical forests. For example, Ledo & Schnitzer [5] found that lianas, which comprised ~35% of the woody species diversity in a Panamanian tropical forest [60,61], were underdispersed (clustered) rather than overdispersed. Thus, Ledo & Schnitzer [5] concluded that, while there was evidence for CNDD for canopy trees, there was little evidence for CNDD for lianas. Similarly, in a Caribbean tropical forest, DeWalt and colleagues [36] found that non-canopy tree woody seedlings (lianas and shrubs) were less likely to suffer negative density dependent mortality than canopy trees. In tropical forests, however, trees commonly represent 65% or more of the woody plant species diversity (e.g., [60,61]), and thus CNDD is still likely a powerful diversity maintenance mechanism. By contrast, CNDD may fail to maintain the majority of species diversity in temperate forests where canopy trees represent a small minority of species [62,63].

In temperate forests, CNDD likely does not occur in isolation. Rather, CNDD and other mechanisms like facilitation, niche specialization, and dispersal limitation likely interact to maintain diversity in these forests. CNDD may be the most important mechanism for the maintenance of tree species diversity even though these other mechanisms are likely to be occurring simultaneously. But for other plant groups, these other mechanisms like facilitation, niche specialization, and dispersal limitation may be more important relative to CNDD. For example, Ledo and Schnitzer [5], found that clumped spatial distributions may be due to niche specialization in lianas, while trees demonstrated overdispersion indicating that CNDD may be more powerful. Similarly, the relative importance of these different mechanisms may change as plants grow. For example, Yao et al. [25] found that CNDD was important for individuals when they were young and small but that topographic and edaphic factors increased in importance with increasing plant age. Similarly, for tree seedlings invading into a grassland, Wright et al. [64] found that smaller tree seedlings benefited from facilitation in high diversity contexts while larger tree seedlings experienced strong competition.

At Powdermill Nature Reserve, a similar scenario where overall diversity is maintained by several mechanisms which simultaneously support diversity but also tradeoff in importance depending on the age/size of individuals and their abiotic context. Trees (and especially the largest trees) may be maintained largely by CNDD; whereas, understory plants may be influenced by a number of different mechanisms. There is evidence that CNDD is a weak mechanism for the maintenance of understory plant diversity, since overdispersion is present when understory plants are small (Fig 2C). However, the lack of overdispersion in larger understory plants indicates that a mechanism (or mechanisms) other than CNDD is a stronger driver of understory plant diversity. Short distance dispersal is often adaptive because site conditions are likely to be the same in the area immediately surrounding a parent plant [65]. Because dispersal syndromes that favor shorter distance dispersal are more common in the understory, mechanisms like niche differentiation that rely on adaptation to specific abiotic factors as found by both Ledo and Schnitzer [5] and Yao et al. [25] may be more important for these understory species.

Canopy trees may be significantly more overdispersed than understory species simply because being tall enables longer distance dispersal. We found higher overdispersion of canopy trees than for understory plants regardless of dispersal mechanism (Fig 4B). That is, tall species were more likely to experience overdispersion whether they were bird or wind dispersed even though bird dispersal may enable more (generally rare) events of very long-distance dispersal [66,67]. Understory plants tend to have universally smaller dispersal kernels regardless of dispersal mechanism because of their smaller stature [53]. Small stature results in fewer seeds dispersed at longer distances—even for bird-dispersed seeds (Fig 4B). The inability to move seeds far away from the parent tree may force understory plants to be better defended against soil pathogens, which appear to be strong agents of CNDD [1,58,68–70]. Furthermore, negative feedback from soil pathogens may be inversely related to light availability (Smith & Reynolds 2015, Jiang et al. 2020) [71,72]. Many understory plants are naturally well defended because of the importance of preserving plant tissue in a low-light environment [73,74]; thus, understory plants may be predisposed to developing greater defenses to pathogens rather than increasing dispersal abilities.

Differences in the level of overdispersion between canopy species and understory species did not appear to be due to the spatial scale of study in spite of our relatively small plot size. If spatial scale had biased our results, we would have expected the spatial point pattern analysis to show little evidence of overdispersion for large canopy trees, but rather a signature indistinguishable from complete spatial random. Furthermore, Zhu et al. [30] demonstrated that when NDD is present it is most likely to be present at the 0–5 m scale and peaks at 5 m (see also [29]). Our results showed a clear spatial signature of overdispersion for our largest individuals. Thus, it seems unlikely that our findings were caused by differences in plant scale. Furthermore, Bagchi & Illian [46] demonstrate that replicated point pattern analysis is significantly more robust to problems of small scale than traditional point pattern analysis.

Conclusions

The intense focus on canopy trees, and in particular on tree seedlings, may bias the current understanding of diversity maintenance in forest ecosystems [25,26]. If we had restricted our sampling to only the smallest understory individuals, we would have concluded that CNDD maintains woody understory plant diversity but not canopy tree diversity. However, examining larger individuals indicated that adult canopy trees became overdispersed as they matured, but that understory plants did not. Zhu and colleagues [18], Detto and colleagues [26] and Yao and colleagues [25] all emphasized similar caution in drawing large-scale conclusions from studies of seedling dynamics for three reasons. First, patterns of seedling mortality often have little effect on broader community and demographic patterns [18]. Second, NDD tends to decrease with ontogeny rather than increase [25]. Finally, studies of NDD at the recruitment level may overestimate NDD due to regression dilution ([26]; but see [4]).

To fully understand the maintenance of plant species diversity, it is necessary to examine spatial patterns across plant sizes, as well as across plant groups that vary in life history strategies. Spatial patterns may be even more complex when considering species that vary more broadly in their life history strategies, such as herbaceous species, which comprise the majority of plant diversity in temperate deciduous forests [63] and are largely neglected with respect to their diversity maintenance [17]. Nevertheless, even by simply dividing the woody plant community into canopy trees and woody understory plants, we demonstrate that CNDD, which appears to maintain canopy tree diversity, may not be strong enough to overcome dispersal limitation and maintain understory woody plant diversity in this temperate forest.

Supporting information

S1 Table. Linear estimates of the relationship between L and distance for all pooled point patterns at Powdermill Nature Reserve.

We report any pooled point pattern as overdispersed if it has a significantly positive slope and any pooled point pattern as clustered if it has a significantly negative slope.

(DOCX)

Click here for additional data file.^{(14.8KB, docx)}

S2 Table. Linear estimates of the relationship between L and distance for all pooled point patterns utilizing degrees of freedom based on the number of points represented by each point pattern rather than the degrees of freedom based on the number of L estimates.

To make a more conservative estimate of significance, we calculated the P value for each pooled point pattern using the standard deviation of the L estimates and the number of points contributing to each point pattern. We used the number of points contributing rather than the number of L estimates because L is calculated 51 times (each 10 cm distance bin) for each individual point pattern resulting in an inflated degrees of freedom for the overall model. We then calculated the t statistic as the slope/standard error and used a T table to find the estimated P value for a two-tailed t test. We report the P value for each T statistic at the closest degrees of freedom on the table to our degrees of freedom that was not greater than the actual degrees of freedom (i.e. for a degrees of freedom of 204, we report the p-value for 200 degrees of freedom). This analysis may be overly conservative because the variance, standard deviation, and standard error are calculated based on the L estimates which have a higher variance (as they are calculated 51 times per point) than the average L estimate for each point.

(DOCX)

Click here for additional data file.^{(16.1KB, docx)}

S3 Table. List of species from Powdermill Nature Reserve.

We classified species using the Flora of North America species descriptions. If a species had an average height of 5 m or higher, we classified it as a canopy species. If a species had an average height of 5 m or lower, we classified it as an understory species. We based our dispersal syndrome on the description of seed morphology.

(DOCX)

Click here for additional data file.^{(15.1KB, docx)}

Acknowledgments

The authors would like to thank Robert Bagchi for helpful comments on the manuscript. The authors would also like to thank Arie Hunt and Joe Strini for field assistance, Jacob Slyder and James Whitacre for GIS and GPS assistance, Cokie Lindsay for administrative and tactical support, and John Wenzel for input on study design and statistical efforts as well as general support. Thanks also to M. Elizabeth Rodriguez Ronderos, Sergio Estrada Villegas, and Sasha Wright for comments on early drafts of this manuscript.

Data Availability

All data and code for this project are now available in my public GitHub repository here: https://github.com/katie-barry44/barry-schnitzer2021PlosOne.

Funding Statement

This work was funded by two Rea Fellowships awarded by the Carnegie Museum of Natural History to K.E.B. Funding for field assistance was provided by a Research Growth Initiative from the University of Wisconsin-Milwaukee to S.A.S. Additional funding for K.E.B. was provided by the Ivy Balsam-Milwaukee Audobon Society Grant.

References

1.Mangan SA, Schnitzer SA, Herre EA, Mack KML, Valencia MC, Sanchez EI, et al. Negative plant–soil feedback predicts tree-species relative abundance in a tropical forest. Nature. 2010;466: 752–755. doi: 10.1038/nature09273 [DOI] [PubMed] [Google Scholar]
2.Johnson DJ, Beaulieu WT, Bever JD, Clay K. Conspecific Negative Density Dependence and Forest Diversity. Science. 2012;336: 904–907. doi: 10.1126/science.1220269 [DOI] [PubMed] [Google Scholar]
3.Comita LS, Queenborough SA, Murphy SJ, Eck JL, Xu K, Krishnadas M, et al. Testing predictions of the Janzen-Connell hypothesis: a meta-analysis of experimental evidence for distance- and density-dependent seed and seedling survival. Gómez-Aparicio L, editor. Journal of Ecology. 2014;102: 845–856. doi: 10.1111/1365-2745.12232 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.LaManna JA, Mangan SA, Myers JA. Conspecific negative density dependence and why its study should not be abandoned. Ecosphere. 2021;12: e03322. 10.1002/ecs2.3322. [DOI] [Google Scholar]
5.Ledo A, Schnitzer SA. Disturbance and clonal reproduction determine liana distribution and maintain liana diversity in a tropical forest. Ecology. 2014;95: 2169–2178. doi: 10.1890/13-1775.1 [DOI] [PubMed] [Google Scholar]
6.Murrell DJ. On the emergent spatial structure of size-structured populations: when does self-thinning lead to a reduction in clustering? Journal of Ecology. 2009;97: 256–266. doi: 10.1111/j.1365-2745.2008.01475.x [DOI] [Google Scholar]
7.LaManna JA, Mangan SA, Alonso A, Bourg NA, Brockelman WY, Bunyavejchewin S, et al. Plant diversity increases with the strength of negative density dependence at the global scale. Science. 2017;356: 1389–1392. doi: 10.1126/science.aam5678 [DOI] [PubMed] [Google Scholar]
8.Janzen DH. Herbivores and the number of tree species in tropical forests. American Naturalist. 1970; 501–528. [Google Scholar]
9.Connell JH. On the role of natural enemies in preventing competitive exclusion in some marine animals and in rain forest trees. Dynamics of populations. 1971;298: 312. [Google Scholar]
10.Anderson TW. Predator responses, prey refuges, and density-dependent mortality of a marine fish. Ecology. 2001;82: 245–257. [Google Scholar]
11.Goldberg DE, Turkington R, Olsvig-Whittaker L, Dyer AR. Density dependence in an annual plant community: variation among life history stages. Ecological Monographs. 2001;71: 423–446. [Google Scholar]
12.Lorenzen K, Enberg K. Density-dependent growth as a key mechanism in the regulation of fish populations: evidence from among-population comparisons. Proceedings of the Royal Society of London B: Biological Sciences. 2002;269: 49–54. doi: 10.1098/rspb.2001.1853 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Petermann JS, Fergus AJF, Turnbull LA, Schmid B. Janzen-Connell effects are widespread and strong enough to maintain diversity in grasslands. Ecology. 2008;89: 2399–2406. doi: 10.1890/07-2056.1 [DOI] [PubMed] [Google Scholar]
14.Comita LS, Muller-Landau HC, Aguilar S, Hubbell SP. Asymmetric density dependence shapes species abundances in a tropical tree community. Science. 2010;329: 330–332. doi: 10.1126/science.1190772 [DOI] [PubMed] [Google Scholar]
15.Schnitzer SA, Klironomos JN, HilleRisLambers J, Kinkel LL, Reich PB, Xiao K, et al. Soil microbes drive the classic plant diversity-productivity pattern. Ecology. 2011;92: 296–303. doi: 10.1890/10-0773.1 [DOI] [PubMed] [Google Scholar]
16.Johnson DJ, Bourg NA, Howe R, McShea WJ, Wolf A, Clay K. Conspecific negative density-dependent mortality and the structure of temperate forests. Ecology. 2014;95: 2493–2503. [Google Scholar]
17.Barry KE, Schnitzer S. Maintenance of Plant Species Diversity in Forest Ecosystems. First. In: Ansari AA, Gill SS, editors. Plant diversity: Monitoring, Assessment, and Conservation. First. UK: CAB International; 2017. [Google Scholar]
18.Zhu K, Woodall CW, Monteiro JVD, Clark JS. Prevalence and strength of density-dependent tree recruitment. Ecology. 2015;96: 2319–2327. doi: 10.1890/14-1780.1 [DOI] [PubMed] [Google Scholar]
19.Hubbell SP. Tree dispersion, abundance, and diversity in a tropical dry forest. Science. 1979;203: 1299–1309. doi: 10.1126/science.203.4387.1299 [DOI] [PubMed] [Google Scholar]
20.Hubbell SP. Seed predation and the coexistence of tree species in tropical forests. Oikos. 1980; 214–229. [Google Scholar]
21.Muller-Landau HC, Adler FR. How seed dispersal affects interactions with specialized natural enemies and their contribution to the maintenance of diversity. Seed dispersal: Theory and its application in a changing world. 2007; 407–446. [Google Scholar]
22.Holt RD. Species Coexistence. In: Levin SA, editor. Encyclopedia of Biodiversity (Second Edition). Waltham: Academic Press; 2013. pp. 667–678. doi: [DOI] [Google Scholar]
23.Packer A, Clay K. Soil pathogens and spatial patterns of seedling mortality in a temperate tree. Nature. 2000;404: 278–281. doi: 10.1038/35005072 [DOI] [PubMed] [Google Scholar]
24.Packer A, Clay K. Development of negative feedback during successive growth cycles of black cherry. Proceedings of the Royal Society of London Series B: Biological Sciences. 2004;271: 317–324. doi: 10.1098/rspb.2003.2583 [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Yao J, Bachelot B, Meng L, Qin J, Zhao X, Zhang C. Abiotic niche partitioning and negative density dependence across multiple life stages in a temperate forest in northeastern China. Journal of Ecology. 2020;108: 1299–1310. 10.1111/1365-2745.13335. [DOI] [Google Scholar]
26.Detto M, Visser MD, Wright SJ, Pacala SW. Bias in the detection of negative density dependence in plant communities. Ecology Letters. 2019;22: 1923–1939. doi: 10.1111/ele.13372 [DOI] [PubMed] [Google Scholar]
27.MacArthur RH, Wilson EO. An equilibrium theory of insular zoogeography. Evolution. 1963; 373–387. [Google Scholar]
28.Condit R, Hubbell SP, Foster RB. Recruitment Near Conspecific Adults and the Maintenance of Tree and Shrub Diversity in a Neotropical Forest. The American Naturalist. 1992;140: 261–286. doi: 10.1086/285412 [DOI] [PubMed] [Google Scholar]
29.Guo Y, Lu Z, Wang Q, Lu J, Xu Y, Meng H, et al. Detecting density dependence from spatial patterns in a heterogeneous subtropical forest of central China. Canadian Journal of Forest Research. 2015. [cited 6 May 2021]. doi: 10.1139/cjfr-2014-0390 [DOI] [Google Scholar]
30.Zhu Y, Mi X, Ren H, Ma K. Density dependence is prevalent in a heterogeneous subtropical forest. Oikos. 2010;119: 109–119. [Google Scholar]
31.Piao T, Comita LS, Jin G, Kim JH. Density dependence across multiple life stages in a temperate old-growth forest of northeast China. Oecologia. 2013;172: 207–217. doi: 10.1007/s00442-012-2481-y [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Yao J, Zhang X, Zhang C, Zhao X, von Gadow K. Effects of density dependence in a temperate forest in northeastern China. Scientific Reports. 2016;6: 32844. doi: 10.1038/srep32844 [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Gilliam FS. The Ecological Significance of the Herbaceous Layer in Temperate Forest Ecosystems. BioScience. 2007;57: 845. doi: 10.1641/B571007 [DOI] [Google Scholar]
34.McCarthy-Neumann S, Kobe RK. Tolerance of soil pathogens co-varies with shade tolerance across species of tropical tree seedlings. Ecology. 2008;89: 1883–1892. doi: 10.1890/07-0211.1 [DOI] [PubMed] [Google Scholar]
35.Kobe RK, Vriesendorp CF. Conspecific density dependence in seedlings varies with species shade tolerance in a wet tropical forest. Ecology letters. 2011;14: 503–510. doi: 10.1111/j.1461-0248.2011.01612.x [DOI] [PubMed] [Google Scholar]
36.DeWalt SJ, Taylor BN, Ickes K. Density-dependent Survival in Seedlings Differs among Woody Life-forms in Tropical Wet Forests of a Caribbean Island. Biotropica. 2015;47: 310–319. doi: 10.1111/btp.12216 [DOI] [Google Scholar]
37.Murphy SJ, Audino LD, Whitacre J, Eck JL, Wenzel JW, Queenborough SA, et al. Species associations structured by environment and land-use history promote beta-diversity in a temperate forest. Ecology. 2015;96: 705–715. doi: 10.1890/14-0695.1 [DOI] [PubMed] [Google Scholar]
38.Zhu Y, Getzin S, Wiegand T, Ren H, Ma K. The relative importance of Janzen-Connell effects in influencing the spatial patterns at the Gutianshan subtropical forest. PloS one. 2013;8: e74560. doi: 10.1371/journal.pone.0074560 [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Flora of North America @ efloras.org. [cited 17 Apr 2012]. Available: http://www.efloras.org/flora_page.aspx?flora_id=1.
40.R Core Team. R: A language and environment for statistical computer. Vienna, Austria: R Foundation for Statistical Computing; 2019. [Google Scholar]
41.Ripley BD. Modelling spatial patterns. Journal of the Royal Statistical Society Series B (Methodological). 1977; 172–212. [Google Scholar]
42.Besag JE. Comments on Ripley’s paper. Journal of the Royal Statistical Society B. 1977;39: 193–195. [Google Scholar]
43.Baddeley A, Turner R. spatstat: An R package for Analyzing Spatial Point Patterns. Journal of Statistical Software. 2005;12: 1–42. [Google Scholar]
44.Baddeley A, Rubak E, Turner R. Spatial Point Patterns: Methodology and Applications with R. Boca Raton, FL, USA: Chapman & Hall/CRC; 2015. Available: https://www.crcpress.com/Spatial-Point-Patterns-Methodology-and-Applications-with-R/Baddeley-Rubak-Turner/9781482210200. [Google Scholar]
45.Brown C, Illian JB, Burslem DFRP. Success of spatial statistics in determining underlying process in simulated plant communities. J Ecol. 2015; n/a-n/a. doi: 10.1111/1365-2745.12493 [DOI] [Google Scholar]
46.Bagchi R, Illian JB. A method for analysing replicated point patterns in ecology. Methods Ecol Evol. 2015;6: 482–490. doi: 10.1111/2041-210X.12335 [DOI] [Google Scholar]
47.Wickham H. Reshaping data with the reshape package. Journal of Statistical Software. 2007;21. [Google Scholar]
48.Aguie B. gridExtra: Miscellaneous Functions for “Grid” Graphics. 2015. doi: 10.1017/S1368980014000378 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Plate T, Heiberger R. abind: Combine Multidimensional Arrays. 2015. [Google Scholar]
50.Wickham H. ggplot2: Elegant graphics for data analysis. New York, N.Y.: Springer; 2009. [Google Scholar]
51.Besag J. Some methods of statistical analysis for spatial data. Bulletin of the International Statistical Institute. 1977;47: 77–92. [Google Scholar]
52.Dale MRT. Spatial pattern analysis in plant ecology. Cambridge, UK: Cambridge University Press; 1999. [Google Scholar]
53.van der Pijl L. Principals of dispersal in higher plants. Berlin, Germany: Springer Verlag; 1982. [Google Scholar]
54.Bates D, Maechler M, Bolker B, Walker S. lme4: Linear mixed effects models using Eigen and S4. 2015. [Google Scholar]
55.Kuznetsova A, Brockhoff PB, Christensen RHB. lmerTest: Tests in Linear Mixed Effects Models. 2015. [Google Scholar]
56.Gray L, He F. Spatial point-pattern analysis for detecting density-dependent competition in a boreal chronosequence of Alberta. Forest Ecology and Management. 2009;259: 98–106. doi: 10.1016/j.foreco.2009.09.048 [DOI] [Google Scholar]
57.Forrister DL, Endara M-J, Younkin GC, Coley PD, Kursar TA. Herbivores as drivers of negative density dependence in tropical forest saplings. Science. 2019;363: 1213–1216. doi: 10.1126/science.aau9460 [DOI] [PubMed] [Google Scholar]
58.Jia S, Wang X, Yuan Z, Lin F, Ye J, Lin G, et al. Tree species traits affect which natural enemies drive the Janzen-Connell effect in a temperate forest. Nat Commun. 2020;11: 286. doi: 10.1038/s41467-019-14140-y [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Antonovics J, Levin DA. The Ecological and Genetic Consequences of Density-Dependent Regulation in Plants. Annual Review of Ecology and Systematics. 1980;11: 411–452. [Google Scholar]
60.Schnitzer SA, Mangan SA, Dalling JW, Baldeck CA, Hubbell SP, Ledo A, et al. Abundance Liana, Diversity, and Distribution on Barro Colorado Island, Panama. Mohan J, editor. PLoS ONE. 2012;7: e52114. doi: 10.1371/journal.pone.0052114 [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Schnitzer S A, Mangan SA, Hubbell SP. The lianas of Barro Colorado Island, Panama. In: Schnitzer S A, Bongers F, Burnham RJ, Putz FE, editors. Ecology of Lianas. UK: John Wiley & Sons; 2015. pp. 76–90. [Google Scholar]
62.Gilliam FS. The herbaceous layer in forests of eastern North America. Oxford University Press; 2014. [Google Scholar]
63.Spicer ME, Mellor H, Carson WP. Seeing beyond the trees: a comparison of tropical and temperate plant growth forms and their vertical distribution. Ecology. 2020;101: e02974. doi: 10.1002/ecy.2974 [DOI] [PubMed] [Google Scholar]
64.Wright A, Schnitzer SA, Reich PB. Daily environmental conditions determine the competition–facilitation balance for plant water status. J Ecol. 2015;103: 648–656. doi: 10.1111/1365-2745.12397 [DOI] [Google Scholar]
65.Bonte D, Hovestadt T, Poethke H-J. Evolution of dispersal polymorphism and local adaptation of dispersal distance in spatially structured landscapes. Oikos. 2010;119: 560–566. 10.1111/j.1600-0706.2009.17943.x. [DOI] [Google Scholar]
66.Nathan R, Muller-Landau HC. Spatial patterns of seed dispersal, their determinants and consequences for recruitment. Trends in Ecology & Evolution. 2000;15: 278–285. doi: 10.1016/s0169-5347(00)01874-7 [DOI] [PubMed] [Google Scholar]
67.Dennis AJ. Seed Dispersal: Theory and Its Application in a Changing World. CABI; 2007. [Google Scholar]
68.Bever JD. Soil community feedback and the coexistence of competitors: conceptual frameworks and empirical tests. New Phytologist. 2003;157: 465–473. doi: 10.1046/j.1469-8137.2003.00714.x [DOI] [PubMed] [Google Scholar]
69.Packer A, Clay K. Soil pathogens and prunus serotina seedling and sapling growth near conspecific trees. Ecology. 2003;84: 108–119. doi: 10.1890/0012-9658(2003)084[0108:SPAPSS]2.0.CO;2 [DOI] [Google Scholar]
70.Kulmatiski A, Beard KH, Stevens JR, Cobbold SM. Plant-soil feedbacks: a meta-analytical review. Ecology Letters. 2008;11: 980–992. doi: 10.1111/j.1461-0248.2008.01209.x [DOI] [PubMed] [Google Scholar]
71.Smith LM, Reynolds HL. Plant–soil feedbacks shift from negative to positive with decreasing light in forest understory species. Ecology. 2015;96: 2523–2532. doi: 10.1890/14-2150.1 [DOI] [PubMed] [Google Scholar]
72.Jiang F, Zhu K, Cadotte MW, Jin G. Tree mycorrhizal type mediates the strength of negative density dependence in temperate forests. Journal of Ecology. 2020;108: 2601–2610. 10.1111/1365-2745.13413 [DOI] [Google Scholar]
73.Coley PD. Herbivory and Defensive Characteristics of Tree Species in a Lowland Tropical Forest. Ecological Monographs. 1983;53: 209–229. doi: 10.2307/1942495 [DOI] [Google Scholar]
74.Coley PD, Bryant JP, Chapin FS III. Resource availability and plant antiherbivore defense. Science (Washington). 1985;230: 895–899. doi: 10.1126/science.230.4728.895 [DOI] [PubMed] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0245639.r001

Decision Letter 0

RunGuo Zang

19 Mar 2021

PONE-D-20-40957

Are we missing the forest for the trees? Conspecific negative density dependence in a temperate deciduous forest

PLOS ONE

Dear Dr. Barry,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

ACADEMIC EDITOR:Please make more clarifications on how to explain the possible reasons of your results

Please ensure that your decision is justified on PLOS ONE’s publication criteria and not, for example, on novelty or perceived impact.

For Lab, Study and Registered Report Protocols: These article types are not expected to include results but may include pilot data.

Please submit your revised manuscript by May 03 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

RunGuo Zang

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

Additional Editor Comments (if provided):

Please make revisions according to the concerns of the reviewers

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: No

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This manuscript studied three conditions that they claim necessary for conspecific negative density dependence (CNDD) to maintain species coexistence by conducting plot surveys in a temperate forest and using spatial point pattern analyses. The study species included different plant growth forms (although results for mid-story trees and lianas were not reported in this manuscript), growth stages, and dispersal modes. Results showed that plants were overdispersed overall, which was a pattern driven by larger growth stages of canopy tree species (“adult canopy trees”) but not understory plants. Because understory plant species can make up the majority of species composition in temperate forest plant diversity, focusing on trees to draw conclusions on CNDD as a mechanism to maintain diversity in forest communities would overestimate its importance where the importance of CNDD in maintaining species diversity.

I liked how this study included growth stages beyond seedlings, which most studies on CNDD are focused on, as I also agree that effects of small seedling mortality may be limited on broader community dynamics. However, there are two major concerns.

First, the framework of this study is not well integrated in the context of existing (a large number of) literature on the topic. It is not a bad idea to test the three conditions (i.e, most individuals will be overdispersed due to CNDD, the degree of CNDD should increase with growth stage due to compounding effects, and CNDD will operate across species with different life history strategies). However, these three conditions are not necessary or sufficient for CNDD to promote species coexistence. This study appears to assume that overdispersion is a result of CNDD (the first condition) and use overdispersion to detect CNDD in the second and third conditions but as one of the key ideas in the manuscript, not sufficient mechanism and rationale connecting the analyses and CNDD are not provided (e.g., see a paper by Gray & He 2009 Forest Ecology and Management). In addition, even without CNDD, one would expect overdispersed patterns in larger individuals because older (and likely larger) individuals are more likely to have lost its true parent trees, which are more likely to be their closest adults (due to initial clumping and dispersal limitation).

In addition, I was wondering why this manuscript never mentioned that (1) most tropical species are clumped and not overdispersed (e.g., Hubbell 1979 Science, which this manuscript cites but I believe miss-cited in line 65; Condit et al. 2000 Science; Armestro et al. 1986 Biotropica, which includes temperate forests); (2) for CNDD to promote species diversity, common species should suffer stronger CNDD than rare species because species diversity is inherently related to species’ rarity (tropical forest diversity is possible by having many rare species and a handful of common species); and (3) there is accumulating evidence (e.g., Bennett et al. 2017 Science; Jiang et al. 2021 Ecology) that the strength and the sign of density dependence can be largely determined by the type of mycorrhizal association plants have. This may also explain less aggregated spatial patterns in temperate forests (e.g., Armestro et al. 1986 Biotropica). Many species included in this study (e.g., Betula, Carpinus, Carya, Fagus, Quercus) are ectomycorrhizal (unlike many tropical species), which often show less CNDD or positive density dependence.

Second concern relates to the sampling method used in this study. Although there is brief rationale about using small plot sizes (Lines 132-141; 289-298), I am really not sure how one can conduct rigorous spatial point pattern analyses using only 16 plots of only 10 m diameter (which is in total 0.13 ha, even when it is across a bigger spatial scale). One cannot even calculate distances beyond 10 m (and much smaller for larger trees) within a plot. Some large trees can have a crown size greater than 10 m but was the forest where the study took place consists of all small trees? Were distances calculated among plots? Using a mean 5 m height as a cutoff of “canopy” tree (lines 566-567) makes me wonder (a shrub can easily be 5+ m tall)... If “adult canopy trees” (line 36) are trees 5-10 m tall, then I do not think “canopy trees” in this manuscript is what most readers would be thinking. In line 135, it cites three papers but none of them used such a small plot sampling scheme and use plots greater than 24 ha. I am guessing this was a side project of another bigger project but for a project trying to cover different growth stages, the study design is flawed. That is why the distance in the figures are up to 6 m, but I am not convinced that any conclusion made at such a small spatial scale can be extrapolated to any dynamics at a larger scale. Perhaps, would that explain some of the inconsistencies in results with other previous studies? Also, at such a small spatial scale, existence of a plot (and out of only 16 plots) that happened to have included a reproductive tree would heavily bias the results (and avoiding them intentionally would also likely bias the included trees to be smaller). Excluding gaps may also bias the results and limit the generality of the findings because species associated with gaps can be rare.

Minor notes

Page 4: This is a field research but details are not provided.

Line 69: What is a “rare species effect”?

Lines 117-146: When was the field survey conducted and over what time frame?

Line 106: How was each growth stage defined? Is it different from the height classes (Lines142-143)? How were these size classes determined?

Line 318: CNDD is known to be strong in grasses; how do you reconcile it you’re your results?

Figure 3: What explains “self-dispersed” not to be the most aggregated compared to other dispersal modes?

Reviewer #2: PLOS ONE

Barry & Schnitzer 2021—Manuscript # PONE-D-20-40957

Overall statement:

This manuscript used spatially explicit plant survey data from a temperate deciduous forest to test conspecific negative density dependence. It is important research which addresses the (lack of!) generality of a leading ecological theory for diversity maintenance. The novelty of the study lies in the system: the vast majority of CNDD studies come from tree data, whereas this study looks at all woody understory species (including vines and shrubs). However, prior to publication in PLOS ONE, I have a few suggestions to clarify the manuscript and put the results into context:

1. Do not over-reach your results in the intro or discussion, because herbaceous plants were not included in the surveys. Your study is a great addition to the CNDD literature because most studies come from trees, but because you did not include herbaceous species (the most species-rich growth form in your study system), I would be careful not to be too broad in calling the analyzed community the whole understory. Your main point/contribution is still true, just modify by saying “woody understory”.

2. Provide justification for your height classes, because they seem somewhat arbitrary to compare among different plant growth-forms (is height a good proxy for ontogeny to compare trees and non-trees?)

3. Overall the discussion could use a little more depth of digging into the mechanisms or the “why” of the results

I also have included comments below for minor changes to improve clarity.

In-line comments and section-specific recommendations:

Abstract:

In-line comments

37 I would be careful here—using that “80% of plant diversity” implies you looked at herbaceous species, which you didn’t. Shrubs make up more like ~10% of the temperate forest vascular plant species, which is still more than trees (~7%) (see Gilliam 2007 BioScience, Spicer et al. 2020 Ecology). I’d temper this argument in the abstract to just make the argument that CNDD has never (? scarcely?) been tested in any growth form other than trees and lianas. You can still easily argue that trees make up a tiny minority of the species in temperate forests.

Introduction:

In-line comments

47 Typo in LaManna’s name in citation

79-89 Rephrase a little unclear here

90 It might be good to guide the reader specifically why growing near a conspecific adult would get worse over time.

95 Be more specific in what you mean by 20% of the community

99-102 Other than the “most species aren’t overstory trees” argument, I think you need to argue why theoretically we expect CNDD to be stronger (or weaker) in non-tree growth forms. You hint at shorter dispersal distances, so would that translate to stronger or weaker NDD? Expand a little more.

Methods

General comments

The only hesitation I have with your methods broadly is your height classes. Assuming you’re using this as a proxy for ontogenic stage, do we know if different growth forms should be grouped in the same height classes? Couldn’t a reproducing adult shrub be <0.5m? And might vine and tree seedlings grow at very different rates? Just wondering if there is a better (but still reasonable) proxy, or if you could divide into ontogenic stages more directly. If not, please justify the height categories (assumedly based on literature from trees), make an argument for why they should be good approximations for shrubs and understory trees too, and make sure to connect the reader to the “why”—will the same height categories be competing?

I think one of your supplementals should be a list of the species with their life-history traits. This would be useful for future studies and to clarify how many species of each category were in the forest.

I also ask for one clarification for the statistical analyses: specify that you put all factors in one model (which is clear by looking at the tables, but less clear in the methods). As written, because there are several “levels” at which the analyses were run: with all plants combined, with individuals separated by size, by growth form, and by dispersal mechanisms, those could each be separate models. One sentence would suffice to explicitly state the variables in the model.

In-line comments

139 Do you have the actual range of accuracy (when taking the GPS points), rather than “up to 10cm”? (Especially if the seedlings were closer together than your less-accurate readings were)

142 In your intro you explicitly said “throughout ontogeny”. Be specific here and say you are using plant size as a proxy for age. Do we know how valid this is for shrubs? Are there good data on how shrub size changes through ontogeny? (See general comments above)

144-145 I like this dispersal syndrome approach to understanding why you see the patterns, but I feel like you missed an opportunity to connect to theory/expectation. Which groups would you predict to have stronger CNDD? How much do these vary within or between growth form groups (so would they be confounded)?

152 I would suggest adding a real quick phrase to justify/explain L (e.g. “for ease of interpretation”)

155 Did the cutoff for removing a species have to be 5 individuals per plot, or total? Were any plots “empty” (no species with >5 individuals/species)? The parenthetical statement makes it sound like there were originally more than 16 plots, but the earlier section says there were 16 originally (I would just take the parenthetical out here if so). If not, how many plots were removed? Or just have a short statement on not analyzing “empty” plots.

157-158 By “species type” do you mean growth form (or life-history strategies)? Stay consistent with terminology or specify your categories somewhere in the methods section. In the introduction (lines 105-106), you specify five types: “shrubs, understory trees, mid-story trees, canopy trees, and lianas”, but just report “canopy trees” versus “understory plants”. How did you categorize them? This is listed in Table S2, but it should more explicitly be referred to in the methods.

161 Couldn’t complete random be a possibility (not likely, but possible)? If so, replace “to ensure” with “to compare to complete random”—isn’t that the null model?

166-188 The authors made a noble attempt to explain these nuanced predictions and justify their interpretation of L(d)-d; it still takes the reader on a bit of a roller coaster. Would it be possible to just put in a supplemental figure that shows a predictions table/figure? This seems so much easier to see rather than imagining from pretty technical prose.

225-227 This is a really good clarification (but maybe belongs in the methods?)

Results

General comments:

Compelling results, interesting, and well-displayed.

Discussion

General comments:

Overall, I wanted the discussion to dig into the mechanisms more, and further explore why we might expect CNDD to be less important for understory plants than overstory trees. What did we learn from the height classes versus the growth form analysis? Might also be interesting to mention the non-native species and just see if they are doing anything different, or discuss expectations for more growth forms (shrubs vs. understory trees versus expectations for lianas or herbs).

Also, why does the shape of the pattern (L(d)-d vs. Distance; Figure 2b) look so different for >5m trees versus the others? Discuss the biological significance/interpretation of the <0.5m trees having zero L(d)-d?

In-line comments

259-273 These two paragraphs seem like too many sentences to say “mechanisms other than CNDD are at play” without actually suggestion what they may be. Expand by suggesting what mechanisms are most likely driving the understory patterns, or eliminate because the next paragraphs get into the mechanisms for trees.

Conclusions

Conclusion provides a nice synthesis of the main results and implications.

In-line comments

305-306 This specific results sentence is not necessary for the conclusion.

Figures

Table 1. Can you put the biological interpretation of a significant L statistic right here in the caption (like in Supplemental Table 1)? It also might help to bold the significant ones.

Figure 2a,b. Even though the colors correspond to 2a, change the labels somewhere on the actual figure to say “overstory” (2b) and “understory” (2c). It also might help interpretation to remind the reader in the figure caption what the biological interpretation is of the 95% CI shaded regions overlapping each other versus overlapping 0 (rather than just say what the conclusion is).

Figure 3b. I would pick another line style for the zero line (or for the canopy-wind) because they are the same.

Table S1: This table would be easier to quick glean the message if the overdispersed or clustered values were emphasized differently (e.g., overdispersed bolded, clustered italicized). As is, the reader has to look at both the t-stat and the p value to interpret.

Table S2: This was only ever referenced once, and not explained at all. Maybe one quick sentence in the methods “we confirmed…with an even more conservative estimate”

Table S2: Were you really able to tell the Carya and Fraxinus species apart at such small life stages? Typo: de-capitalize Americana for Fraxinus, and isn’t the common name white ash?

566 An average height of 5m or higher from your dataset, or reported in the Flora of North America? Would be good to mention because one could also argue for max height, rather than average, being a good metric of over- versus understory status (can it ever break through the understory, versus does the species on average?).

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2021 Jul 15;16(7):e0245639. doi: 10.1371/journal.pone.0245639.r002

Author response to Decision Letter 0

14 Jun 2021

We have also submitted this information as a .docx with the rest of our materials where the formatting makes it a bit easier to see our response to individual comments.

PONE-D-20-40957

Are we missing the forest for the trees? Conspecific negative density dependence in a temperate deciduous forest

PLOS ONE

Dear Dr. Barry,

ACADEMIC EDITOR:Please make more clarifications on how to explain the possible reasons of your results

Please ensure that your decision is justified on PLOS ONE’s publication criteria and not, for example, on novelty or perceived impact.

For Lab, Study and Registered Report Protocols: These article types are not expected to include results but may include pilot data.

Please include the following items when submitting your revised manuscript:

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

We look forward to receiving your revised manuscript.

Kind regards,

RunGuo Zang

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Additional Editor Comments (if provided):

Please make revisions according to the concerns of the reviewers

Many thanks again to the editor for their constructive oversight over this process. We have numbered the comments of the reviewers for ease of response and respond to them directly below the comment.

Many thanks to the review for their excellent summary of this work and for their kind words regarding the inclusion of different growth stages. We have now numbered their comments for ease of response and respond to them in bold.

1. First, the framework of this study is not well integrated in the context of existing (a large number of) literature on the topic. It is not a bad idea to test the three conditions (i.e, most individuals will be overdispersed due to CNDD, the degree of CNDD should increase with growth stage due to compounding effects, and CNDD will operate across species with different life history strategies). However, these three conditions are not necessary or sufficient for CNDD to promote species coexistence. This study appears to assume that overdispersion is a result of CNDD (the first condition) and use overdispersion to detect CNDD in the second and third conditions but as one of the key ideas in the manuscript, not sufficient mechanism and rationale connecting the analyses and CNDD are not provided (e.g., see a paper by Gray & He 2009 Forest Ecology and Management). In addition, even without CNDD, one would expect overdispersed patterns in larger individuals because older (and likely larger) individuals are more likely to have lost its true parent trees, which are more likely to be their closest adults (due to initial clumping and dispersal limitation).

First, we thank the reviewer for agreeing that it is a good idea to test the three conditions responsible for CNDD. The reviewer raises a good point about other potential drivers of overdispersion of larger stems and we have added a paragraph to the Discussion stating that the increase in the probability of over dispersion with plant size could have several potential explanations (Lines 237 to 239) which read:

“However, we cannot rule out the possibility that trees may be more likely to be overdispersed with size simply because the larger (and presumably older) the tree the greater the probability of mortality for the parent (which is often nearby) [59].”

Nonetheless, we appreciate that the reviewer acknowledges the benefits of examining multiple plant size classes. In addition, we now state that these three conditions are consistent with CNDD; however, our approach goes further than other studies that typically test the over dispersion component of CNDD.

2. In addition, I was wondering why this manuscript never mentioned that (1) most tropical species are clumped and not overdispersed (e.g., Hubbell 1979 Science, which this manuscript cites but I believe miss-cited in line 65; Condit et al. 2000 Science; Armestro et al. 1986 Biotropica, which includes temperate forests); (2) for CNDD to promote species diversity, common species should suffer stronger CNDD than rare species because species diversity is inherently related to species’ rarity (tropical forest diversity is possible by having many rare species and a handful of common species); and (3) there is accumulating evidence (e.g., Bennett et al. 2017 Science; Jiang et al. 2021 Ecology) that the strength and the sign of density dependence can be largely determined by the type of mycorrhizal association plants have. This may also explain less aggregated spatial patterns in temperate forests (e.g., Armestro et al. 1986 Biotropica). Many species included in this study (e.g., Betula, Carpinus, Carya, Fagus, Quercus) are ectomycorrhizal (unlike many tropical species), which often show less CNDD or positive density dependence.

We also address the reviewer’s comment that tree species distribution is mostly clumped. The spatial pattern of tree distribution is likely related to scale. Several studies that show that tropical tree seedling and sapling performance and distribution are over dispersed in relation to conspecifics (e.g., Comita et al. 2010, Ledo & Schnitzer 2014), but that at larger scales these same species may appear clumped. We now clarify this on lines 81 to 85) which read:

“Currently, the evidence for CNDD beyond the seed to seedling transition is mixed. For example, a study by Yao et al. [25] found that CNDD decreased with increasing tree ontogeny in a temperate forest. In fact, many species in both temperate and tropical forests do not have an overdispersed distribution [19,28]. By contrast, Guo et al. [29] found that 75% of tree species demonstrated CNDD as adults in subtropical forests (see also [30–32]).”

Another point raised by the reviewer is that common species suffer more from CNDD than rare species. However, several influential studies have shown the opposite: that adult species frequency is negatively correlated with the strength of CNDD (e.g., Comita et al. 2010, Mangan et al. 2010, Johnson et al. 2012). That is, rare species suffer more from proximity to conspecific adults than common species. We agree with the reviewer that, theoretically, common species should have strongest CNDD; nonetheless, empirical data do not support this expectation.

3. Second concern relates to the sampling method used in this study. Although there is brief rationale about using small plot sizes (Lines 132-141; 289-298), I am really not sure how one can conduct rigorous spatial point pattern analyses using only 16 plots of only 10 m diameter (which is in total 0.13 ha, even when it is across a bigger spatial scale). One cannot even calculate distances beyond 10 m (and much smaller for larger trees) within a plot. Some large trees can have a crown size greater than 10 m but was the forest where the study took place consists of all small trees? Were distances calculated among plots? Using a mean 5 m height as a cutoff of “canopy” tree (lines 566-567) makes me wonder (a shrub can easily be 5+ m tall)... If “adult canopy trees” (line 36) are trees 5-10 m tall, then I do not think “canopy trees” in this manuscript is what most readers would be thinking. In line 135, it cites three papers but none of them used such a small plot sampling scheme and use plots greater than 24 ha. I am guessing this was a side project of another bigger project but for a project trying to cover different growth stages, the study design is flawed. That is why the distance in the figures are up to 6 m, but I am not convinced that any conclusion made at such a small spatial scale can be extrapolated to any dynamics at a larger scale. Perhaps, would that explain some of the inconsistencies in results with other previous studies? Also, at such a small spatial scale, existence of a plot (and out of only 16 plots) that happened to have included a reproductive tree would heavily bias the results (and avoiding them intentionally would also likely bias the included trees to be smaller). Excluding gaps may also bias the results and limit the generality of the findings because species associated with gaps can be rare.

Second, we appreciate the reviewer’s concerns about the size and extent of our study. We believe that our findings are robust based on our sampling scheme. However, we now directly address the sampling issues raised by the reviewer and we discuss how the size of the study could have influenced our results (300-309) which read:

“Differences in the level of overdispersion between canopy species and understory species did not appear to be due to the spatial scale of study in spite of our relatively small plot size. If spatial scale had biased our results, we would have expected the spatial point pattern analysis to show little evidence of overdispersion for large canopy trees, but rather a signature indistinguishable from complete spatial random. Furthermore, Zhu et al. [30] demonstrated that when NDD is present it is most likely to be present at the 0-5 m scale and peaks at 5 m (see also [29]). Our results showed a clear spatial signature of overdispersion for our largest individuals. Thus, it seems unlikely that our findings were caused by differences in plant scale. Furthermore, Bagchi & Illian [46] demonstrate that replicated point pattern analysis is significantly more robust to problems of small scale than traditional point pattern analysis.”

Overall, our study was large enough for the results to be consistent with CNDD and we believe that a larger sample size would likely not change our results or interpretations. However, in addition to addressing these concerns in the manuscript, we have also tempered the interpretation of our findings.

Minor notes

Page 4: This is a field research but details are not provided.

We could not be quite sure what the reviewer was referring to. We have now clarified on line 104 that this research was done in the field.

Line 69: What is a “rare species effect”?

We have rephrased this to now say that “CNDD will not theoretically benefit rare species…” (on line 59 of the revised manuscript) which we hope clarifies what we mean by a rare species effect.

Lines 117-146: When was the field survey conducted and over what time frame?

We have now clarified this on line 127, we established these plots in May and June of 2014.

Line 106: How was each growth stage defined? Is it different from the height classes (Lines142-143)? How were these size classes determined?

The use of height classes is now clarified on lines 138 to 142.

Line 318: CNDD is known to be strong in grasses; how do you reconcile it you’re your results?

We have clarified in this version of our paper that our results refer only to woody plants in temperate deciduous forests. They would therefore not be in conflict with results from grasslands.

Figure 3: What explains “self-dispersed” not to be the most aggregated compared to other dispersal modes?

This is a great question! The most likely explanation is that our self-dispersed species managed to get their seeds farther away than plants that were dispersed by some types of animals. Hamamelis virginiana may not be typical of other self-dispersed species because it uses a ballistic method of dispersal so seeds may travel further than expected for self-dispersed species.

Reviewer #2:

First, we would like to extend our many thanks to the reviewer for their overall positive thoughts on the manuscript. We found their comments to be positive, constructive, and detailed and feel that they have contributed to a much stronger manuscript upon resubmission.

However, prior to publication in PLOS ONE, I have a few suggestions to clarify the manuscript and put the results into context:

We have now tried to clarify our contribution throughout the abstract, introduction, and discussion. These revisions are small but impactful. For example, throughout we’ve emphasized that we are speaking about the woody understory rather than the “understory” as we previously wrote.

Thanks to the reviewer for pointing out that this was unclear. We use height classes as a proxy for both relative age (assuming that woody plants get taller as they get older) and also their position in the forest (i.e. are they a in the canopy or in the understory). We use height in this way because it provides a different type of information from just understory vs. overstory or different dispersal syndromes. A small individual of an overstory species likely has a smaller seed shadow than a taller individual of the same species. Because we believe dispersal to be so closely intertwined with CNDD, we felt that height was likely an important factor. We have now clarified this in two places in the manuscript.

A.) In the introduction on lines 89 to 94.

B.) More in depth in the methods on lines 138 to 142 where it now states: “We use these height classes as a proxy for both relative age (assuming that plants get taller as they get older) and position within the forest. Individuals that are shorter are less likely to be able to disperse seeds farther away than individuals that are taller even if they belong to the same species and have the same dispersal syndrome.”

3. Overall the discussion could use a little more depth of digging into the mechanisms or the “why” of the results

Thanks to the reviewer for this comment as we believe it significantly improves the discussion of this paper. We have revised the discussion to better discuss the “why” of our results. This is especially evident in the two paragraphs highlighted by the reviewer in their specific comments on the discussion. These two paragraphs have now been significantly revised in two ways:

1. We’ve removed some of the redundant language from these paragraphs to make the discussion of alternative mechanisms a bit more streamlined.

2. We’ve added more specific discussion of what other mechanisms may be more relevant.

These paragraphs now read (on lines 257 to 284):

“In temperate forests, CNDD likely does not occur in isolation. Rather, CNDD and other mechanisms like facilitation, niche specialization, and dispersal limitation likely interact to maintain diversity in these forests. CNDD may be the most important mechanism for the maintenance of tree species diversity even though these other mechanisms are likely to be occurring simultaneously. But for other plant groups, these other mechanisms like facilitation, niche specialization, and dispersal limitation may be more important relative to CNDD. For example, Ledo and Schnitzer [5], found that clumped spatial distributions may be due to niche specialization in lianas, while trees demonstrated overdispersion indicating that CNDD may be more powerful. Similarly, the relative importance of these different mechanisms may change as plants grow. For example, Yao et al. [25] found that CNDD was important for individuals when they were young and small but that topographic and edaphic factors increased in importance with increasing plant age. Similarly, for tree seedlings invading into a grassland, Wright et al. [64] found that smaller tree seedlings benefited from facilitation in high diversity contexts while larger tree seedlings experienced strong competition.

At Powdermill Nature Reserve, a similar scenario where overall diversity is maintained by several mechanisms which simultaneously support diversity but also tradeoff in importance depending on the age/size of individuals and their abiotic context. Trees (and especially the largest trees) may be maintained largely by CNDD; whereas, understory plants may be influenced by a number of different mechanisms. There is evidence that CNDD is a weak mechanism for the maintenance of understory plant diversity, since overdispersion is present when understory plants are small (Fig. 2c). However, the lack of overdispersion in larger understory plants indicates that a mechanism (or mechanisms) other than CNDD is a stronger driver of understory plant diversity. Short distance dispersal is often adaptive because site conditions are likely to be the same in the area immediately surrounding a parent plant [65]. Because dispersal syndromes that favor shorter distance dispersal are more common in the understory, mechanisms like niche differentiation that rely on adaptation to specific abiotic factors as found by both Ledo and Schnitzer [5] and Yao et al. [25] may be more important for these understory species.”

Abstract

1. Line 37 I would be careful here—using that “80% of plant diversity” implies you looked at herbaceous species, which you didn’t. Shrubs make up more like ~10% of the temperate forest vascular plant species, which is still more than trees (~7%) (see Gilliam 2007 BioScience, Spicer et al. 2020 Ecology). I’d temper this argument in the abstract to just make the argument that CNDD has never (? scarcely?) been tested in any growth form other than trees and lianas. You can still easily argue that trees make up a tiny minority of the species in temperate forests.

We’ve now removed this statement and temper this argument throughout. A good example of this can be found on line 32.

Introduction

1. 47 Typo in LaManna’s name in citation

This is now fixed.

2. 79-89 Rephrase a little unclear here

We have now rephrased this paragraph and removed this statement.

We also revised this partially to accommodate reviewer #2’s major concern #1 as well as reviewer #1’s concern #1.

3. 90 It might be good to guide the reader specifically why growing near a conspecific adult would get worse over time.

We have now added this information – this information is now included on lines 86-88.

4. 95 Be more specific in what you mean by 20% of the community

We have changed this statement also in response to reviewer #2’s comment #1 to say 7% since we are referring to only canopy trees (line 89).

5. 99-102 Other than the “most species aren’t overstory trees” argument, I think you need to argue why theoretically we expect CNDD to be stronger (or weaker) in non-tree growth forms. You hint at shorter dispersal distances, so would that translate to stronger or weaker NDD? Expand a little more.

This is now clarified in this paragraph on lines 92 to 98 where it reads:

“By contrast, understory plants (including understory woody species) represent a larger share of diversity but have a lower capacity for long distance dispersal due to their relatively short stature and position in forest understory. Furthermore, few understory plant species have dispersal syndromes that favor long distance dispersal [33]. Many understory species are gravity dispersed while the majority of temperate canopy trees are wind dispersed. Thus, the strength of CNDD may interact with plant dispersal syndrome.”

Methods

General comments

We respond to this issue in full in our response to reviewer #2 comment #2. Thanks so much again to the reviewer for highlighting this issue.

1. I think one of your supplementals should be a list of the species with their life-history traits. This would be useful for future studies and to clarify how many species of each category were in the forest.

This information can be found in table S3.

2. I also ask for one clarification for the statistical analyses: specify that you put all factors in one model (which is clear by looking at the tables, but less clear in the methods). As written, because there are several “levels” at which the analyses were run: with all plants combined, with individuals separated by size, by growth form, and by dispersal mechanisms, those could each be separate models. One sentence would suffice to explicitly state the variables in the model.

We have now briefly clarified this on line 190 of the methods.

3. 139 Do you have the actual range of accuracy (when taking the GPS points), rather than “up to 10cm”? (Especially if the seedlings were closer together than your less-accurate readings were)

Unfortunately, we do not have this information.

4. 142 In your intro you explicitly said “throughout ontogeny”. Be specific here and say you are using plant size as a proxy for age. Do we know how valid this is for shrubs? Are there good data on how shrub size changes through ontogeny? (See general comments above)

We have now clarified this (as detailed above in our response to reviewer #2 comment #1) on lines 138 to 142.

5. 144-145 I like this dispersal syndrome approach to understanding why you see the patterns, but I feel like you missed an opportunity to connect to theory/expectation. Which groups would you predict to have stronger CNDD? How much do these vary within or between growth form groups (so would they be confounded)?

We have tried to clarify this throughout the manuscript. However, we find this a bit of a tough line to walk since our hypothesis seeks to test the generality of CNDD and our original expectation was that CNDD would be common even in understory species. Thus, while we do provide some brief expectations in the introduction in this version of the manuscript as suggested here by Reviewer #2 (see lines 92 to 98), we have largely expounded upon this in the discussion (also to address Reviewer #2’s major comment #3.

6. 152 I would suggest adding a real quick phrase to justify/explain L (e.g. “for ease of interpretation”)

This phrase is now added on line 150.

7. 155 Did the cutoff for removing a species have to be 5 individuals per plot, or total? Were any plots “empty” (no species with >5 individuals/species)? The parenthetical statement makes it sound like there were originally more than 16 plots, but the earlier section says there were 16 originally (I would just take the parenthetical out here if so). If not, how many plots were removed? Or just have a short statement on not analyzing “empty” plots.

This sentence is now clarified on lines 153-155.

8. 157-158 By “species type” do you mean growth form (or life-history strategies)? Stay consistent with terminology or specify your categories somewhere in the methods section. In the introduction (lines 105-106), you specify five types: “shrubs, understory trees, mid-story trees, canopy trees, and lianas”, but just report “canopy trees” versus “understory plants”. How did you categorize them? This is listed in Table S2, but it should more explicitly be referred to in the methods.

This sentence has now been clarified and we now use growth form throughout to refer only to canopy vs. understory woody plants.

9. 161 Couldn’t complete random be a possibility (not likely, but possible)? If so, replace “to ensure” with “to compare to complete random”—isn’t that the null model?

This is totally correct. We have now changed this in the manuscript on line 162-163.

10. 166-188 The authors made a noble attempt to explain these nuanced predictions and justify their interpretation of L(d)-d; it still takes the reader on a bit of a roller coaster. Would it be possible to just put in a supplemental figure that shows a predictions table/figure? This seems so much easier to see rather than imagining from pretty technical prose.

This is a great suggestion and we include now a figure (Figure 1 a,b,c) that does exactly this in the new version of the manuscript. This figure shows a basic conceptual distribution as well as a hypothetical overdispersed and underdispersed distribution.

11. 225-227 This is a really good clarification (but maybe belongs in the methods?)

Thanks for this suggestion, we’ve now moved it to the methods on lines 157-160.

Results

General comments:

Compelling results, interesting, and well-displayed.

Thanks so much to the reviewer for including these positive comments as well as the more critical ones!

Discussion

General comments:

1. Overall, I wanted the discussion to dig into the mechanisms more, and further explore why we might expect CNDD to be less important for understory plants than overstory trees. What did we learn from the height classes versus the growth form analysis? Might also be interesting to mention the non-native species and just see if they are doing anything different, or discuss expectations for more growth forms (shrubs vs. understory trees versus expectations for lianas or herbs).

2. Also, why does the shape of the pattern (L(d)-d vs. Distance; Figure 2b) look so different for >5m trees versus the others? Discuss the biological significance/interpretation of the <0.5m trees having zero L(d)-d?

In-line comments

3. 259-273 These two paragraphs seem like too many sentences to say “mechanisms other than CNDD are at play” without actually suggestion what they may be. Expand by suggesting what mechanisms are most likely driving the understory patterns, or eliminate because the next paragraphs get into the mechanisms for trees.

Thanks to the reviewer for pointing this out. We have now revised this paragraph (see our response to reviewer #2 major comment #3.

Conclusions

Conclusion provides a nice synthesis of the main results and implications.

In-line comments

305-306 This specific results sentence is not necessary for the conclusion.

We’ve now removed this statement.

Figures

Table 1. Can you put the biological interpretation of a significant L statistic right here in the caption (like in Supplemental Table 1)? It also might help to bold the significant ones.

Thanks to the reviewer for pointing this out, we have accordingly modified the figure and table captions.

Figure 3b. I would pick another line style for the zero line (or for the canopy-wind) because they are the same.

We’ve tried this a few ways and found it difficult to interpret with a modified line style (since the line styles get increasingly complicated). We’ve tried to clarify this in the figure caption instead and hope this helps.

While we include these results in the supplement, our interpretation of over- vs. underdispersal depends on the interaction between the results in this table and the figures themselves. For example, if a pattern was significantly overdispersed according to the statistical result but overlapped completely with complete spatial random, we did not consider this a significant result. Our concern with bolding and clustering these results is that it may oversimplify this interpretation and cause more confusion.

Table S2: This was only ever referenced once, and not explained at all. Maybe one quick sentence in the methods “we confirmed…with an even more conservative estimate”

This has now been added on lines 192-194.

Table S2: Were you really able to tell the Carya and Fraxinus species apart at such small life stages? Typo: de-capitalize Americana for Fraxinus, and isn’t the common name white ash?

All individuals were >10cm in height and we found that this largely excluded very small seedlings of both Carya and Fraxinus but that we were able to determine these at the species level with a fair degree of accuracy at this height. We consulted with a local botanist when we were unsure of plant identities. We’ve also fixed the typos in this table.

This has now been clarified in the figure caption. We used the average height of the species as reported in the Flora of North America. We agree that max. height may be a potentially better metric, however – this would not change the species that we identify as understory vs. overstory as none of the understory species break through to the canopy in eastern deciduous forests.

Attachment

Submitted filename: Barry+Schnitzer-PONE-R1-Response-9.6.2021.docx

Click here for additional data file.^{(36KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0245639.r003

Decision Letter 1

RunGuo Zang

18 Jun 2021

Are we missing the forest for the trees? Conspecific negative density dependence in a temperate deciduous forest

PONE-D-20-40957R1

Dear Dr. Barry,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

RunGuo Zang

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

PLoS One. doi: 10.1371/journal.pone.0245639.r004

Acceptance letter

RunGuo Zang

1 Jul 2021

PONE-D-20-40957R1

Are we missing the forest for the trees? Conspecific negative density dependence in a temperate deciduous forest

Dear Dr. Barry:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor RunGuo Zang

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Table. Linear estimates of the relationship between L and distance for all pooled point patterns at Powdermill Nature Reserve.

We report any pooled point pattern as overdispersed if it has a significantly positive slope and any pooled point pattern as clustered if it has a significantly negative slope.

(DOCX)

Click here for additional data file.^{(14.8KB, docx)}

(DOCX)

Click here for additional data file.^{(16.1KB, docx)}

S3 Table. List of species from Powdermill Nature Reserve.

(DOCX)

Click here for additional data file.^{(15.1KB, docx)}

Attachment

Submitted filename: Barry+Schnitzer-PONE-R1-Response-9.6.2021.docx

Click here for additional data file.^{(36KB, docx)}

Data Availability Statement

All data and code for this project are now available in my public GitHub repository here: https://github.com/katie-barry44/barry-schnitzer2021PlosOne.

[pone.0245639.ref001] 1.Mangan SA, Schnitzer SA, Herre EA, Mack KML, Valencia MC, Sanchez EI, et al. Negative plant–soil feedback predicts tree-species relative abundance in a tropical forest. Nature. 2010;466: 752–755. doi: 10.1038/nature09273 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref002] 2.Johnson DJ, Beaulieu WT, Bever JD, Clay K. Conspecific Negative Density Dependence and Forest Diversity. Science. 2012;336: 904–907. doi: 10.1126/science.1220269 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref003] 3.Comita LS, Queenborough SA, Murphy SJ, Eck JL, Xu K, Krishnadas M, et al. Testing predictions of the Janzen-Connell hypothesis: a meta-analysis of experimental evidence for distance- and density-dependent seed and seedling survival. Gómez-Aparicio L, editor. Journal of Ecology. 2014;102: 845–856. doi: 10.1111/1365-2745.12232 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0245639.ref004] 4.LaManna JA, Mangan SA, Myers JA. Conspecific negative density dependence and why its study should not be abandoned. Ecosphere. 2021;12: e03322. 10.1002/ecs2.3322. [DOI] [Google Scholar]

[pone.0245639.ref005] 5.Ledo A, Schnitzer SA. Disturbance and clonal reproduction determine liana distribution and maintain liana diversity in a tropical forest. Ecology. 2014;95: 2169–2178. doi: 10.1890/13-1775.1 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref006] 6.Murrell DJ. On the emergent spatial structure of size-structured populations: when does self-thinning lead to a reduction in clustering? Journal of Ecology. 2009;97: 256–266. doi: 10.1111/j.1365-2745.2008.01475.x [DOI] [Google Scholar]

[pone.0245639.ref007] 7.LaManna JA, Mangan SA, Alonso A, Bourg NA, Brockelman WY, Bunyavejchewin S, et al. Plant diversity increases with the strength of negative density dependence at the global scale. Science. 2017;356: 1389–1392. doi: 10.1126/science.aam5678 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref008] 8.Janzen DH. Herbivores and the number of tree species in tropical forests. American Naturalist. 1970; 501–528. [Google Scholar]

[pone.0245639.ref009] 9.Connell JH. On the role of natural enemies in preventing competitive exclusion in some marine animals and in rain forest trees. Dynamics of populations. 1971;298: 312. [Google Scholar]

[pone.0245639.ref010] 10.Anderson TW. Predator responses, prey refuges, and density-dependent mortality of a marine fish. Ecology. 2001;82: 245–257. [Google Scholar]

[pone.0245639.ref011] 11.Goldberg DE, Turkington R, Olsvig-Whittaker L, Dyer AR. Density dependence in an annual plant community: variation among life history stages. Ecological Monographs. 2001;71: 423–446. [Google Scholar]

[pone.0245639.ref012] 12.Lorenzen K, Enberg K. Density-dependent growth as a key mechanism in the regulation of fish populations: evidence from among-population comparisons. Proceedings of the Royal Society of London B: Biological Sciences. 2002;269: 49–54. doi: 10.1098/rspb.2001.1853 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0245639.ref013] 13.Petermann JS, Fergus AJF, Turnbull LA, Schmid B. Janzen-Connell effects are widespread and strong enough to maintain diversity in grasslands. Ecology. 2008;89: 2399–2406. doi: 10.1890/07-2056.1 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref014] 14.Comita LS, Muller-Landau HC, Aguilar S, Hubbell SP. Asymmetric density dependence shapes species abundances in a tropical tree community. Science. 2010;329: 330–332. doi: 10.1126/science.1190772 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref015] 15.Schnitzer SA, Klironomos JN, HilleRisLambers J, Kinkel LL, Reich PB, Xiao K, et al. Soil microbes drive the classic plant diversity-productivity pattern. Ecology. 2011;92: 296–303. doi: 10.1890/10-0773.1 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref016] 16.Johnson DJ, Bourg NA, Howe R, McShea WJ, Wolf A, Clay K. Conspecific negative density-dependent mortality and the structure of temperate forests. Ecology. 2014;95: 2493–2503. [Google Scholar]

[pone.0245639.ref017] 17.Barry KE, Schnitzer S. Maintenance of Plant Species Diversity in Forest Ecosystems. First. In: Ansari AA, Gill SS, editors. Plant diversity: Monitoring, Assessment, and Conservation. First. UK: CAB International; 2017. [Google Scholar]

[pone.0245639.ref018] 18.Zhu K, Woodall CW, Monteiro JVD, Clark JS. Prevalence and strength of density-dependent tree recruitment. Ecology. 2015;96: 2319–2327. doi: 10.1890/14-1780.1 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref019] 19.Hubbell SP. Tree dispersion, abundance, and diversity in a tropical dry forest. Science. 1979;203: 1299–1309. doi: 10.1126/science.203.4387.1299 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref020] 20.Hubbell SP. Seed predation and the coexistence of tree species in tropical forests. Oikos. 1980; 214–229. [Google Scholar]

[pone.0245639.ref021] 21.Muller-Landau HC, Adler FR. How seed dispersal affects interactions with specialized natural enemies and their contribution to the maintenance of diversity. Seed dispersal: Theory and its application in a changing world. 2007; 407–446. [Google Scholar]

[pone.0245639.ref022] 22.Holt RD. Species Coexistence. In: Levin SA, editor. Encyclopedia of Biodiversity (Second Edition). Waltham: Academic Press; 2013. pp. 667–678. doi: [DOI] [Google Scholar]

[pone.0245639.ref023] 23.Packer A, Clay K. Soil pathogens and spatial patterns of seedling mortality in a temperate tree. Nature. 2000;404: 278–281. doi: 10.1038/35005072 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref024] 24.Packer A, Clay K. Development of negative feedback during successive growth cycles of black cherry. Proceedings of the Royal Society of London Series B: Biological Sciences. 2004;271: 317–324. doi: 10.1098/rspb.2003.2583 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0245639.ref025] 25.Yao J, Bachelot B, Meng L, Qin J, Zhao X, Zhang C. Abiotic niche partitioning and negative density dependence across multiple life stages in a temperate forest in northeastern China. Journal of Ecology. 2020;108: 1299–1310. 10.1111/1365-2745.13335. [DOI] [Google Scholar]

[pone.0245639.ref026] 26.Detto M, Visser MD, Wright SJ, Pacala SW. Bias in the detection of negative density dependence in plant communities. Ecology Letters. 2019;22: 1923–1939. doi: 10.1111/ele.13372 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref027] 27.MacArthur RH, Wilson EO. An equilibrium theory of insular zoogeography. Evolution. 1963; 373–387. [Google Scholar]

[pone.0245639.ref028] 28.Condit R, Hubbell SP, Foster RB. Recruitment Near Conspecific Adults and the Maintenance of Tree and Shrub Diversity in a Neotropical Forest. The American Naturalist. 1992;140: 261–286. doi: 10.1086/285412 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref029] 29.Guo Y, Lu Z, Wang Q, Lu J, Xu Y, Meng H, et al. Detecting density dependence from spatial patterns in a heterogeneous subtropical forest of central China. Canadian Journal of Forest Research. 2015. [cited 6 May 2021]. doi: 10.1139/cjfr-2014-0390 [DOI] [Google Scholar]

[pone.0245639.ref030] 30.Zhu Y, Mi X, Ren H, Ma K. Density dependence is prevalent in a heterogeneous subtropical forest. Oikos. 2010;119: 109–119. [Google Scholar]

[pone.0245639.ref031] 31.Piao T, Comita LS, Jin G, Kim JH. Density dependence across multiple life stages in a temperate old-growth forest of northeast China. Oecologia. 2013;172: 207–217. doi: 10.1007/s00442-012-2481-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0245639.ref032] 32.Yao J, Zhang X, Zhang C, Zhao X, von Gadow K. Effects of density dependence in a temperate forest in northeastern China. Scientific Reports. 2016;6: 32844. doi: 10.1038/srep32844 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0245639.ref033] 33.Gilliam FS. The Ecological Significance of the Herbaceous Layer in Temperate Forest Ecosystems. BioScience. 2007;57: 845. doi: 10.1641/B571007 [DOI] [Google Scholar]

[pone.0245639.ref034] 34.McCarthy-Neumann S, Kobe RK. Tolerance of soil pathogens co-varies with shade tolerance across species of tropical tree seedlings. Ecology. 2008;89: 1883–1892. doi: 10.1890/07-0211.1 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref035] 35.Kobe RK, Vriesendorp CF. Conspecific density dependence in seedlings varies with species shade tolerance in a wet tropical forest. Ecology letters. 2011;14: 503–510. doi: 10.1111/j.1461-0248.2011.01612.x [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref036] 36.DeWalt SJ, Taylor BN, Ickes K. Density-dependent Survival in Seedlings Differs among Woody Life-forms in Tropical Wet Forests of a Caribbean Island. Biotropica. 2015;47: 310–319. doi: 10.1111/btp.12216 [DOI] [Google Scholar]

[pone.0245639.ref037] 37.Murphy SJ, Audino LD, Whitacre J, Eck JL, Wenzel JW, Queenborough SA, et al. Species associations structured by environment and land-use history promote beta-diversity in a temperate forest. Ecology. 2015;96: 705–715. doi: 10.1890/14-0695.1 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref038] 38.Zhu Y, Getzin S, Wiegand T, Ren H, Ma K. The relative importance of Janzen-Connell effects in influencing the spatial patterns at the Gutianshan subtropical forest. PloS one. 2013;8: e74560. doi: 10.1371/journal.pone.0074560 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0245639.ref039] 39.Flora of North America @ efloras.org. [cited 17 Apr 2012]. Available: http://www.efloras.org/flora_page.aspx?flora_id=1.

[pone.0245639.ref040] 40.R Core Team. R: A language and environment for statistical computer. Vienna, Austria: R Foundation for Statistical Computing; 2019. [Google Scholar]

[pone.0245639.ref041] 41.Ripley BD. Modelling spatial patterns. Journal of the Royal Statistical Society Series B (Methodological). 1977; 172–212. [Google Scholar]

[pone.0245639.ref042] 42.Besag JE. Comments on Ripley’s paper. Journal of the Royal Statistical Society B. 1977;39: 193–195. [Google Scholar]

[pone.0245639.ref043] 43.Baddeley A, Turner R. spatstat: An R package for Analyzing Spatial Point Patterns. Journal of Statistical Software. 2005;12: 1–42. [Google Scholar]

[pone.0245639.ref044] 44.Baddeley A, Rubak E, Turner R. Spatial Point Patterns: Methodology and Applications with R. Boca Raton, FL, USA: Chapman & Hall/CRC; 2015. Available: https://www.crcpress.com/Spatial-Point-Patterns-Methodology-and-Applications-with-R/Baddeley-Rubak-Turner/9781482210200. [Google Scholar]

[pone.0245639.ref045] 45.Brown C, Illian JB, Burslem DFRP. Success of spatial statistics in determining underlying process in simulated plant communities. J Ecol. 2015; n/a-n/a. doi: 10.1111/1365-2745.12493 [DOI] [Google Scholar]

[pone.0245639.ref046] 46.Bagchi R, Illian JB. A method for analysing replicated point patterns in ecology. Methods Ecol Evol. 2015;6: 482–490. doi: 10.1111/2041-210X.12335 [DOI] [Google Scholar]

[pone.0245639.ref047] 47.Wickham H. Reshaping data with the reshape package. Journal of Statistical Software. 2007;21. [Google Scholar]

[pone.0245639.ref048] 48.Aguie B. gridExtra: Miscellaneous Functions for “Grid” Graphics. 2015. doi: 10.1017/S1368980014000378 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0245639.ref049] 49.Plate T, Heiberger R. abind: Combine Multidimensional Arrays. 2015. [Google Scholar]

[pone.0245639.ref050] 50.Wickham H. ggplot2: Elegant graphics for data analysis. New York, N.Y.: Springer; 2009. [Google Scholar]

[pone.0245639.ref051] 51.Besag J. Some methods of statistical analysis for spatial data. Bulletin of the International Statistical Institute. 1977;47: 77–92. [Google Scholar]

[pone.0245639.ref052] 52.Dale MRT. Spatial pattern analysis in plant ecology. Cambridge, UK: Cambridge University Press; 1999. [Google Scholar]

[pone.0245639.ref053] 53.van der Pijl L. Principals of dispersal in higher plants. Berlin, Germany: Springer Verlag; 1982. [Google Scholar]

[pone.0245639.ref054] 54.Bates D, Maechler M, Bolker B, Walker S. lme4: Linear mixed effects models using Eigen and S4. 2015. [Google Scholar]

[pone.0245639.ref055] 55.Kuznetsova A, Brockhoff PB, Christensen RHB. lmerTest: Tests in Linear Mixed Effects Models. 2015. [Google Scholar]

[pone.0245639.ref056] 56.Gray L, He F. Spatial point-pattern analysis for detecting density-dependent competition in a boreal chronosequence of Alberta. Forest Ecology and Management. 2009;259: 98–106. doi: 10.1016/j.foreco.2009.09.048 [DOI] [Google Scholar]

[pone.0245639.ref057] 57.Forrister DL, Endara M-J, Younkin GC, Coley PD, Kursar TA. Herbivores as drivers of negative density dependence in tropical forest saplings. Science. 2019;363: 1213–1216. doi: 10.1126/science.aau9460 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref058] 58.Jia S, Wang X, Yuan Z, Lin F, Ye J, Lin G, et al. Tree species traits affect which natural enemies drive the Janzen-Connell effect in a temperate forest. Nat Commun. 2020;11: 286. doi: 10.1038/s41467-019-14140-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0245639.ref059] 59.Antonovics J, Levin DA. The Ecological and Genetic Consequences of Density-Dependent Regulation in Plants. Annual Review of Ecology and Systematics. 1980;11: 411–452. [Google Scholar]

[pone.0245639.ref060] 60.Schnitzer SA, Mangan SA, Dalling JW, Baldeck CA, Hubbell SP, Ledo A, et al. Abundance Liana, Diversity, and Distribution on Barro Colorado Island, Panama. Mohan J, editor. PLoS ONE. 2012;7: e52114. doi: 10.1371/journal.pone.0052114 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0245639.ref061] 61.Schnitzer S A, Mangan SA, Hubbell SP. The lianas of Barro Colorado Island, Panama. In: Schnitzer S A, Bongers F, Burnham RJ, Putz FE, editors. Ecology of Lianas. UK: John Wiley & Sons; 2015. pp. 76–90. [Google Scholar]

[pone.0245639.ref062] 62.Gilliam FS. The herbaceous layer in forests of eastern North America. Oxford University Press; 2014. [Google Scholar]

[pone.0245639.ref063] 63.Spicer ME, Mellor H, Carson WP. Seeing beyond the trees: a comparison of tropical and temperate plant growth forms and their vertical distribution. Ecology. 2020;101: e02974. doi: 10.1002/ecy.2974 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref064] 64.Wright A, Schnitzer SA, Reich PB. Daily environmental conditions determine the competition–facilitation balance for plant water status. J Ecol. 2015;103: 648–656. doi: 10.1111/1365-2745.12397 [DOI] [Google Scholar]

[pone.0245639.ref065] 65.Bonte D, Hovestadt T, Poethke H-J. Evolution of dispersal polymorphism and local adaptation of dispersal distance in spatially structured landscapes. Oikos. 2010;119: 560–566. 10.1111/j.1600-0706.2009.17943.x. [DOI] [Google Scholar]

[pone.0245639.ref066] 66.Nathan R, Muller-Landau HC. Spatial patterns of seed dispersal, their determinants and consequences for recruitment. Trends in Ecology & Evolution. 2000;15: 278–285. doi: 10.1016/s0169-5347(00)01874-7 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref067] 67.Dennis AJ. Seed Dispersal: Theory and Its Application in a Changing World. CABI; 2007. [Google Scholar]

[pone.0245639.ref068] 68.Bever JD. Soil community feedback and the coexistence of competitors: conceptual frameworks and empirical tests. New Phytologist. 2003;157: 465–473. doi: 10.1046/j.1469-8137.2003.00714.x [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref069] 69.Packer A, Clay K. Soil pathogens and prunus serotina seedling and sapling growth near conspecific trees. Ecology. 2003;84: 108–119. doi: 10.1890/0012-9658(2003)084[0108:SPAPSS]2.0.CO;2 [DOI] [Google Scholar]

[pone.0245639.ref070] 70.Kulmatiski A, Beard KH, Stevens JR, Cobbold SM. Plant-soil feedbacks: a meta-analytical review. Ecology Letters. 2008;11: 980–992. doi: 10.1111/j.1461-0248.2008.01209.x [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref071] 71.Smith LM, Reynolds HL. Plant–soil feedbacks shift from negative to positive with decreasing light in forest understory species. Ecology. 2015;96: 2523–2532. doi: 10.1890/14-2150.1 [DOI] [PubMed] [Google Scholar]

[pone.0245639.ref072] 72.Jiang F, Zhu K, Cadotte MW, Jin G. Tree mycorrhizal type mediates the strength of negative density dependence in temperate forests. Journal of Ecology. 2020;108: 2601–2610. 10.1111/1365-2745.13413 [DOI] [Google Scholar]

[pone.0245639.ref073] 73.Coley PD. Herbivory and Defensive Characteristics of Tree Species in a Lowland Tropical Forest. Ecological Monographs. 1983;53: 209–229. doi: 10.2307/1942495 [DOI] [Google Scholar]

[pone.0245639.ref074] 74.Coley PD, Bryant JP, Chapin FS III. Resource availability and plant antiherbivore defense. Science (Washington). 1985;230: 895–899. doi: 10.1126/science.230.4728.895 [DOI] [PubMed] [Google Scholar]

PERMALINK

Are we missing the forest for the trees? Conspecific negative density dependence in a temperate deciduous forest

Kathryn E Barry

Stefan A Schnitzer

Roles

Abstract

Introduction

Materials and methods

Study site

Plot establishment and plant census

Data analysis

Fig 1. Conceptual representation of L(d)-d and representative spatial patterns.

Table 1. Results of mixed effects linear model to calculate L statistic significance at Powdermill Nature Reserve in Southwestern Pennsylvania for all comparisons of all individuals >10 cm height.

Results

Fig 2. Pooled Besag’s L statistic across distance from spatial point pattern analysis for the full community of woody plants >10 cm in height at Powdermill Nature Reserve in Southwestern Pennsylvania.

Fig 3. Pooled Besag’s L statistic across distance from spatial point pattern analysis for woody plants >10 cm in height at Powdermill Nature Reserve in Southwestern Pennsylvania separated by growth form.

Fig 4. Pooled Besag’s L statistic across distance from spatial point pattern analysis of the woody plant community stratified by dispersal mechanism and plant type at Powdermill Nature Reserve in southwestern Pennsylvania.

Discussion

Conclusions

Supporting information

Acknowledgments

Data Availability

Funding Statement

References

Decision Letter 0

RunGuo Zang

Roles

Author response to Decision Letter 0

Decision Letter 1

RunGuo Zang

Roles

Acceptance letter

RunGuo Zang

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases