Survival of the weakest: increased frond mechanical strength in a wave-swept kelp inhibits self-pruning and increases whole-plant mortality


Corresponding author. E-mail:


  1. Organisms' ability to withstand the physical forces of their environment is a key determinant of their success. Mechanical performance of organisms is often associated with the properties of the tissues that compose them. In mechanically stressful habitats, intraspecific variation in tissue properties may result in differential survivorship and enable natural selection to act on material performance.
  2. We tested the hypothesis that tissue mechanical properties affect survivorship (a fitness component) of the perennial kelp, Egregia menziesii, in a mechanically stressful, wave-swept intertidal habitat.
  3. We measured intraspecific variation in strength and flexibility in 38 Emenziesii and tracked their survivorship in the field over the winter storm season to determine whether variation in mechanical properties led to differential survivorship.
  4. Significant interindividual variation was found in most mechanical properties, including strength and flexibility. Individuals with increased flexibility and decreased strength were more likely to survive the duration of our study, although this effect was more pronounced in individuals with smaller holdfasts. Increased frond strength was also associated with a reduction in self-thinning, potentially explaining the observed increase in whole-plant mortality with increasing frond strength.
  5. Results from this study demonstrate that variation in tissue mechanical properties among conspecifics can influence survivorship and this may have important evolutionary implications.


Species distributional patterns are determined by their ability to thrive under a range of biotic and abiotic conditions. One often overlooked, albeit crucial, element of environmental variation is the physical forces organisms must endure and organisms exhibit an exquisite diversity of form to meet these challenges. For example, many trees must withstand the violent forces of seasonal storms (Spatz & Bruechert 2004), vertebrate tendons must withstand stresses associated with creating movement, and spider silk must absorb the energy of rapidly moving prey, while adhering to the prey's surface (Agnarsson & Blackledge 2009). Of all the mechanically challenging environments on earth, the wave-swept shore may be one of the most inhospitable owing to extreme mechanical stressors. Organisms within this zone must endure the forces of crashing waves, which can be 100 times greater, for the same-sized organism, than in any reported wind storm, every 5–10 s (Denny & Gaylord 2002). Nonetheless, the wave-swept intertidal hosts a tremendous diversity of plants and animals that exhibit highly adapted forms to meet these challenges.

In order for intertidal organisms to prevent becoming beach wrack, their tissues must remain stronger than the forces they experience. Some species accomplish this by strengthening their tissues (Lowell, Markham & Mann 1991; Martone 2007) while others create flexible tissues that can reconfigure in flow to generate more streamlined shapes (Koehl 1984; Boller & Carrington 2006). Flexibility is often thought of as a prerequisite for life along wave-swept shorelines (Harder et al. 2004) because tissue flexibility dramatically decreases drag forces experienced and most intertidal plants are markedly more flexible than terrestrial plants. Whether individuals create strong and/or flexible tissues, their hydrodynamic performance is largely dependent upon the properties of their tissues (Boller & Carrington 2007; Demes et al. 2011). Therefore, tissue mechanical properties are thought to be under strong selective pressures along high-energy shorelines, where wave-induced damage may be the largest cause of mortality of sessile organisms.

Despite the wealth of research suggesting that tissue mechanical properties dictate performance of organisms in the wave-swept intertidal (e.g. Koehl & Wainwright 1977; Koehl 1984; Denny 1988), there is a paucity of data on how tissues' mechanical properties impact whole-organism survival and reproductive success. Many studies have compared mechanical properties among species (e.g. Harder et al. 2004; Boller & Carrington 2007; Demes et al. 2011) and speculated about the fitness consequences of between-species variation in mechanical design. For instance, Dudgeon & Johnson (1992) related differences in mechanical properties to differential survivorship in two competing species of macroalgae, suggesting that tissues' components can shape differential survival among species. However, whether or how variation in tissue mechanical traits among individuals within a site affects whole-plant survivorship has not been considered explicitly. This deficit persists despite the broadly accepted view that evolution shapes traits by selecting on intraspecific trait variants (Darwin 1859).

In order for evolution to shape organisms' material properties, there must first be heritable within-species variation on which selection can act. Intraspecific variation has been documented in many species of intertidal macroalgae in two contexts: differences among populations and differences within populations. Studies on local adaptation of tissue mechanical properties have demonstrated differences in mechanical strength and stiffness between sites (Johnson & Koehl 1994; Blanchette, Miner & Gaines 2002). Such variation, however, may be due to either genetic divergence between populations or simply plasticity. Within-site variation in mechanical properties has been documented along exposure gradients (Kitzes & Denny 2005) and through laboratory culture studies of individuals from the same site under different conditions (Kraemer & Chapman 1991). Such within-site variation has often been attributed to plasticity, where individuals produce different material properties based on the hydrodynamic cues they receive during development. However, within-site variation in mechanical properties can also arise as the result of underlying genetic variation among individuals. If such variation results in differential performance, it is possible for selection to act on tissue mechanical properties. Studies considering whether or how contemporary selection acts on standing intraspecific variation in mechanical properties are virtually absent.

In this study, we used Egregia menziesii, a common kelp species along exposed coastlines, to test (i) whether intraspecific variation in frond mechanical traits was repeatable; and (ii) whether variation in mechanical properties among conspecifics resulted in differential survivorship. Because this species can live for several years, its holdfast must withstand the largest hydrodynamic forces imposed by winter swells to avoid dislodgement. Black (1976) showed that grazing by the limpet Acmaea incessa increased survival of E. menziesii individuals by pruning fronds (likely as a consequence of decreasing frond breaking strength), resulting in smaller plants during periods of winter swell. Because pruning seems to be related to frond breaking strength, it is also possible that intraspecific variation in mechanical strength, in the absence of herbivores, results in differential pruning and therefore survivorship. Specifically, we predicted that plants with lower frond breaking strength would be more likely to self-prune, thereby decreasing the probability of whole-plant dislodgement (i.e. holdfast failure) during winter storms.

Materials and methods

Study Species and Site

Egregia menziesii (Turner) JE Areschoug (Fig. 1) is a perennial kelp (Laminariales, Phaeophyceae) common at exposed sites from Alaska to Baja California. Adult E. menziesii sporophytes typically consist of ~25 fronds emerging from a single holdfast. In September 2010, 38 E. menziesii individuals with at least 20 fronds were tagged (nondestructively by attaching a foam label through holdfast haptera with a cable tie) at Botanical Beach (48°31′33.46N, 124°26′50.22W), an exposed rocky promontory located near Port Renfrew, British Columbia, Canada. Each month, the number of fronds was recorded as well as whether or not the holdfast had become dislodged. Because spatial variation in hydrodynamic forces could affect survivorship, this study was performed within a small area (~150 m2) along a homogeneous, flat intertidal bench. Maximum wave velocities were characterized to determine whether there were spatial trends in maximum wave forces by randomly placing 10 spring scale dynamometers (Bell & Denny 1994) throughout the study site. The relative positions of all plants and dynamometers were measured to the nearest centimetre to test for spatial autocorrelation.

Figure 1.

Egregia menziesii. Starting at left, then clockwise – close-up photo of a holdfast firmly attached to rock (photo courtesy of Tess Grainger), a mature individual (thinned for photographic purposes) splayed out on a dock to show size (scale = 50 cm), and a close-up of fronds exposed at low tide showing frond structure: blades and pneumatocysts (air bladders) are born on both sides of a central frond axis (rachis).

Determination of Mechanical Properties

Sample collection and processing

In October 2010, five randomly selected fronds from each of the 40 tagged individuals were harvested and transported in seawater from the collection site to Friday Harbor Laboratories for biomechanical analyses. Upon arrival, fronds were placed in flow-through seawater tables until tested. All specimens were tested within 48 h of collection. To ensure that all tissues tested were roughly the same age, the distal-most 11 cm of each rachis, just below the terminal lamina (growth in this species occurs at the rachis apices), was used in mechanical tests. Tissues were then cut into standardized working sections to control localized strain while reducing the likelihood of biases from clamp-induced damage (Mach 2009; Demes et al. 2011). Sample width and length were held constant by the standardized working section, but thickness of each sample was measured to the nearest 0.1 mm using digital calipers before tensile tests. Working sections were then secured, via pneumatic clamps at 90 psi, to a tensometer machine (5565, Instron, Norwood, MA, USA). The Instron strained samples at a rate of 10 mm min−1 and recorded the resisting force (N) every second until tissue failure. Tensile tests were performed in air on wet tissue. Any tissue samples that failed near the attachment clamps were discarded from the analysis.

Calculation of mechanical properties

Strain was calculated as the change in length over the initial length (engineer's strain). Tissue stress was calculated as force (N) per initial cross-sectional area (m2) and reported in MPa. Tensile stiffness (modulus of elasticity, E) was calculated as the slope of stress vs. strain curves. Because the nature of such curves for kelp tissues is r-shaped, stiffness was calculated from the initial linear phase. Breaking stress and breaking stain were taken as the last values before failure occurred. Whole frond breaking force was determined by multiplying breaking stress of a test section by the cross-sectional area of the rachis from which it was taken. Bending (flexural) stiffness was calculated as the product of initial tensile stiffness (E) and the second moment of area (I), which was estimated for rachis tissue using the equation for a beam, I = (1/12)*width*thickness3 (Vogel 2003). Holdfast attachment area of individuals was calculated with the equation for area of a circle using the diameter of the holdfast, which was measured to the nearest centimetre.

Statistical Analyses

Because none of the mechanical properties met the assumption of homogeneity of variances for anova, Welch's anova (Welch 1951) was used to test for repeatability in mechanical properties among individuals. For individuals that did survive the winter, linear regression analysis was used to determine whether mean frond breaking force (N) significantly influenced self-pruning (proportion of fronds retained). Binary logistic regression was used to determine whether mean mechanical properties and holdfast attachment area were associated with whole-plant survivorship. Reverse stepwise (log likelihood ratio) multiple logistic regression was then used to assess the relative importance of holdfast attachment area and mechanical properties in predicting survivorship. Spatial autocorrelation of maximum wave velocity and whole-plant survivorship was tested using Spatial Analysis in Macroecology (SAM), v4.0. Other statistical analyses were implemented in SYSTAT 13 and were considered significant with α = 0.05 and P < 0.05.


Significant differences among individuals were detected in all tissue mechanical properties except breaking strain (Table 1). Linear regression analysis revealed that proportion of initial fronds retained throughout the winter was significantly (n = 25, R2 = 0.218, P = 0.033) positively related to mean rachis breaking force (Fig. 2).

Table 1. Summary of frond mechanical properties and their effects on whole-plant survivorship. Logistic regression values are resultant from binary logistic regression analyses on each factor individually. Bold values indicate statistical significance (P < 0.05)
Mechanical propertyRange of means r Welch's anovaLogistic regression
d.f.F-ratioP-value r 2 P-value
Breaking strain12.3–26.90.1837,44.11.4160.1340.0020.827
Breaking stress (MPa)1.4–3.90.4837, 43.92.904 <0.001 0.0220.429
Tensile stiffness (MPa)4.1–24.60.3937, 41.03.069 <0.001 0.0750.14
Flexural stiffness (μN*m2)8.0–94.40.5937, 40.72.78 <0.001 0.177 0.022
Frond breaking force (N)9.1–46.20.3937, 42.756.36 <0.001 0.212 0.011
Holdfast attachment area (cm2)19.6–188.60.384 <0.001
Figure 2.

Proportion of initial fronds retained through the winter as a function of an individuals' mean frond (rachis) breaking force (N) and fitted linear model. Individuals with stronger fronds kept a greater proportion of their fronds (pruned less).

Binary logistic regression revealed that holdfast attachment area, mean frond breaking force and mean frond flexural stiffness were significantly associated with whole-plant survival (Fig. 3). We failed to detect significant associations with other tissue mechanical properties and survivorship (Table 1). Pearson's correlation analysis failed to detect relationships between holdfast attachment area and frond breaking force (P > 0.4) but did find a positive correlation between flexural stiffness and mean frond breaking force (n = 38, r = 0.353, P = 0.027). Because flexural stiffness and frond breaking force were correlated, only frond breaking force, holdfast attachment area and their interaction were included in the overall model predicting survivorship; frond breaking force was chosen over flexural stiffness because of its higher R2 value (Table 1).

Figure 3.

Survivorship (1 = survival, 0 = dislodgement) as a function of A) holdfast attachment area, B) an individuals' mean frond flexural stiffness and C) an individuals' mean frond (or rachis) breaking force (N) and fitted logistic models.

The full model including frond breaking force, holdfast attachment area and their interaction was significant (P < 0.001) and explained 60% of the variability in whole-plant survivorship. However, removal of the holdfast attachment area term did not significantly change the model's predictive power, resulting in a reduced model containing only mean frond breaking force (P = 0.008) and the interaction between mean frond breaking force and holdfast attachment area (P = 0.018). This reduced model was significant (P < 0.001) and also explained 60% of the variability in whole-plant survivorship. The nature of the significant interaction was such that individuals with larger holdfasts could survive even with unrealistically high frond breaking forces, while individuals with smaller holdfasts survive only with weaker fronds (i.e. more susceptible to self-pruning) (Fig. 4). Further removal of either term from the reduced model resulted in significantly diminished predictive power (P < 0.001). We were unable to detect spatial trends in either maximum wave velocities or whole-plant survivorship (Fig. 5) at any of the spatial scales analysed (≫ 0.05).

Figure 4.

A graphical representation of how mean frond breaking force and holdfast attachment area interact to influence survivorship. Empty circles represent individuals that were dislodged, while filled circles represent individuals that perenniated. The dashed line (discriminant function) represents the axis along which survivorship is differentiated; individuals in the grey area (below the line) are likely to survive, while individuals above this line are at high risk of dislodgement.

Figure 5.

Results from correlograms testing spatial autocorrelation of maximum wave velocity (black bars) and survivorship (grey bars). Reported values are Moran's I values and were not significant (P ≫ 0.05) for both variables at all distance classes.


Understanding how the traits of organisms are honed to meet the challenges of their environment is a perennial goal in organismal biology. Within the wave-swept intertidal, organisms must endure tremendous forces at regular intervals for the duration of their existence, and intuitively, these forces have imposed strong selection pressures on the form and function of these organisms. Although there is considerable interest in characterizing the performance consequences of mechanical properties in marine macrophytes (e.g. drag, flexibility), remarkably few studies have tested the effects of mechanical properties on more direct fitness metrics (e.g. survival, fecundity). In the present study, we document an association between individual E. menziesii's material properties and their persistence. However, the relationship was not entirely intuitive. Rather than increased strength conferring a survival advantage, it was diminished strength that was positively associated with individuals' survival probability. Our data suggest this advantage is conferred in the form of self-pruning, where weaker individuals lose more fronds, which reduces their drag and diminishes their probability of being completely dislodged (Blanchette 1997). However, the importance of frond mechanical properties in predicting whole-plant survivorship was diminished in individuals with larger holdfasts (Fig. 4). Our data represent the first evidence of variation in material properties among individuals within the same site explaining differences in survivorship of marine macrophytes.

Given the marked survival differences among individuals possessing different material properties, we ask ‘Why does intraspecific variation in material properties persist within the population?’ Or alternatively, ‘why do not all individuals possess weaker fronds?’ If survival is the key determinant of individuals' fitness in this system, theory predicts that strong selection should rapidly eliminate inferior trait variants in mechanical properties. However, survival is not the only performance component in marine macrophytes. In fact, the advantage gained by weaker individuals in the form of self-pruning must inevitably come at a cost of reduced productivity. Or, in other words, individuals with a high propensity to self-prune must suffer reduced biomass and productivity as portions of the organism are torn off over time. Thus, we propose trait variation is maintained by a broadly applicable life-history trade-off: individuals may either invest in their persistence (via self-pruning) at the cost of reduced fecundity, or adopt a high risk boom-and-bust strategy, where individuals enjoy high reproductive success but absorb the cost of reduced survivorship. It follows that the benefits of either strategy would vary spatially and temporally; variation among sites and among years in maximum wave velocity could conceivably mediate this trade-off and maintain variation in mechanical properties within E. menziesii populations.

Another mechanism that might generate variation in material properties is condition dependence. In the present study, the extent to which whole-plant survivorship was dependent upon frond mechanical traits was dictated by the size of the individual's holdfast (Fig. 4). The significance of this interaction term suggests a complex selection surface, which may be partially responsible for maintaining variation in frond mechanical traits. For instance, individuals with larger holdfasts may be able to sustain the costs of more (and therefore stronger) fronds. Larger holdfasts could arise from genetic underpinnings or could also result from desirable local environmental factors that enhance productivity and allow greater energy allocation to holdfasts.

Within-population variation in mechanical traits could also arise through plasticity whereby individuals may adopt material properties based on the cues of their environment, or as a mere by-product of variation in resource acquisition. We attempted to minimize the potential for plasticity to have resulted in the observed trends by using a small, homogeneous area and testing for spatial trends in survivorship, mechanical traits and maximum wave forces. However, observed differences among individuals may still be due to either underlying genetic variation, environmental cues or some combination limiting our ability to make inferences about whether or not natural selection is occurring on material properties in this alga. At present, we lack heritability estimates of material properties in E. menziesii, or indeed in any marine macrophyte and, instead, present repeatability, which is often used as a crude proxy for heritability in organisms that cannot be bred in laboratories (e.g. highly exposed kelp species) (Boake 1989; Falconer & McKay 1996; Husak 2006; Pruitt 2010).

Our data reveal significant, repeatable variation in tissue mechanical properties among conspecifics at the same site and that such differences are correlated with differential performance. Interestingly, the only mechanical properties that influenced survivorship (frond breaking force and flexural stiffness) are both composite properties dependent upon both tissue material properties and the amount of tissue present. Martone (2007) highlighted that intertidal seaweeds can increase breaking force by strengthening their tissues or increasing amount of tissue. Because kelps (Laminariales) are so large and happen to be a particularly morphological variable group (e.g. Koehl et al. 2008; Demes, Graham & Suskiewicz 2009), much of their mechanical performance may be related to their morphology (Koehl & Alberte 1988; Johnson & Koehl 1994). This highlights that while the materials intertidal organisms' tissues are made of are important to their performance (Boller & Carrington 2007; Demes et al. 2011), anatomical and morphological strategies may also be important components of mechanical performance (Harley & Bertness 1996; Demes et al. 2011).


In this study, we documented an association between intraspecific variation in mechanical properties and survivorship in E. menziesii, where decreased breaking force was associated with greater survivorship for individuals (especially for those with smaller holdfasts). This finding reaffirms the (often-assumed) notion that organisms' struggle with the physical forces of their environments can shape ecological and evolutionary processes. Documenting links between intraspecific variation in mechanical traits and survivorship is an important piece in the bridge between biomechanical and evolutionary/ecological research, because it directly links trait performance with variation in an individual's fitness. We argue that studies that explicitly consider the effects of mechanical traits on fitness are a vital, but largely missing, component in mechanical models and that biomechanical research could benefit from a more thorough integration with evolutionary ecology research.


We are indebted to Laura Anderson, William Iles, Jennifer Jorve, Rebecca Martone and Jocelyn Nelson for help with field work, Matthew George for help with mechanical testing set-up, and Mary O'Connor and Rob DeWreede for thoughtful discussion throughout the course of this study. This work was funded by a Stephen and Ruth Wainwright fellowship to Kyle Demes and National Science Foundation Grants OCE-0752523 and EF-1041213. Authorization to work in the Botanical Beach protected area was provided by the Canadian Ministry of the Environment (permit number 104149).