254 (2000) 189–209
www.elsevier.nl / locate / jembe
Variations in holdfast attachment mechanics with
developmental stage, substratum-type, season, and
wave-exposure for the intertidal kelp species Hedophyllum sessile
(C. Agardh) Setchell
a ,* a
Kristen L.D. Milligan , Robert E. DeWreede
a
Department of Botany, The University of British Columbia, Vancouver, B.C. Canada V6T 1Z4
Received 24 January 2000; received in revised form 20 April 2000; accepted 11 August 2000
Abstract
Biomechanical models that describe physical and biological interactions on wave-exposed shores typically assume that a species’ attachment properties are similar between seasons and sites. We tested this assumption using Hedophyllum sessile to investigate how macroalgal biomechani-cal attachment properties vary with developmental stage, substratum-type, season, and wave-exposure. Hedophyllum sessile is an intertidal kelp species that is able to survive in wave-exposed areas in the Northeast Pacific. For both juveniles and adults, holdfast attachment force and strength were measured at a wave-exposed and wave-protected site in Barkley Sound, British Columbia, Canada. Substratum and wave-exposure effects on attachment properties were tested in juvenile populations. Adult populations were sampled prior to (in July 1996) and after (in November 1996) a series of storms. Site and seasonal wave-exposure effects on attachment properties were tested in these adult populations. Comparisons to known attachment properties of other temperate macroalgal species were also made. Causes for these patterns are discussed but were not isolated in these studies. Juveniles’ attachment properties differed on different substrata types and between wave-exposures, with the highest attachment forces and the most attached juveniles in articulated
22
coralline algal turfs. Adult attachment is firm (|100 N), but relatively weak (|0.07 MN m ).
Adult attachment did not vary with site wave-exposure, but there was a shift within each site to more resistant holdfasts after a series of early winter storms. Seasonal increases in storm swells correlated to more thallus tattering and selected against large, loose holdfasts. The data presented here suggest that results from holdfast attachment field studies in one season cannot be extrapolated to another due to a complex set of dynamics. This is the first documentation of seasonal patterns in macroalgal attachment properties. 2000 Elsevier Science B.V. All rights reserved.
*Corresponding author. Present address: Clean Ocean Action, P.O. Box 505, Sandy Hook, NJ 07732, USA. Tel.: 11-732-872-0111; fax:11-732-872-8041.
E-mail address: [email protected] (K.L.D. Milligan).
Keywords: Biomechanics; Holdfast; Kelp; Macroalgal ecology; Wave-exposure
1. Introduction
Recent studies on mechanisms of wave-induced mortality in macroalgal populations have demonstrated that biomechanical approaches can explain species distributions and morphological patterns (Gaylord et al., 1994). For example, Shaughnessy et al. (1996) demonstrated that differences in distribution of two closely related species (Mazzaella
linearis and M. splendens, Rhodophyta) in relation to wave-exposure could be predicted by a biomechanical survivorship model. Many of the studies to date focus on variations in thallus attributes, such as tissue mechanics (Koehl and Wainwright, 1977; Biedka et al., 1987; Denny et al., 1989) and thallus shape (Koehl, 1984; Carrington, 1990; Dudgeon and Johnson, 1992), which may resist and minimize wave-induced stresses, and attachment forces that resist wave stress (Carrington, 1990; Friedland and Denny, 1995).
In general, survival in wave-exposed sites is dependent on the total forces transferred from the wave to the thallus attachment and the resistance force of the attachment. If the total hydrodynamically-induced force exceeds the attachment force, then the alga or part of the alga will be torn from the substratum and dislodged (Carrington, 1990). The majority of field measurements of attachment forces resulted in thallus breaks above the holdfast at either the stipe or the stipe / blade junction (Carrington, 1990; Dudgeon and Johnson, 1992; Shaughnessy et al., 1996). This results in partial thallus loss, with the holdfast or holdfast and stipe remaining for potential regrowth. In this case, biomass loss before reproduction reduces the number of propagules available for dispersal and recruitment. In cases of holdfast failure, the whole thallus is dislodged, and space is opened for recolonization.
Despite increased attention towards understanding effects of wave force on intertidal organisms (e.g. Denny, 1988, 1995), there is a lack of empirical studies documenting the variation in attachment properties among and within species, and across gradients such as wave-exposures, seasons, and substratum types. Past model approaches to predict algal survival along wave-exposure gradients have been based on assumptions that a species’ attachment properties do not significantly vary with season or site (e.g. Gaylord et al., 1994; Denny, 1995). However, it is unknown whether or not these assumptions represent natural populations. Biomechanical attributes, such as tissue locations most prone to breakage, forces to break, and tissue strengths, that are observed in the field reflect the complex interactions between intrinsic (such as evolved tissue traits) and extrinsic (for example, herbivore and wave damage) factors which directly affect individual survival in wave-exposed sites. These factors may or may not result in similar attachment properties between seasons or across wave-exposure gradients.
A species’ evolved tissue mechanics contribute to ways in which it resists physical damage. For example, to increase attachment force, either stronger (strength is measured as force per unit area) tissue or increased cross sectional area of break junction is needed. For example, Mazzaella splendens, M. linearis, (Shaughnessy et al., 1996) and
Mastocarpus papillatus (Carrington, 1990) increase resistance to breakage by increasing
Combinations of herbivore- and wave-induced damage to a thallus, such as flaws and cracks in stipes, blades, or holdfasts, are common, causing failure-prone locations (Koehl and Wainwright, 1977; Denny et al., 1989; DeWreede et al., 1992; Tegner et al., 1995). Depending on the severity of wounding from extrinsic factors, the thallus will ultimately break at locations different from those that occur in non-damaged thalli (Biedka et al., 1987).
Holdfast failure has not been studied in detail even though it is a commonly observed occurrence, especially in some kelp species (Koehl and Wainwright, 1977; Tegner et al., 1995). Holdfast loosening is a natural progression in kelp maturity, resulting from individual haptera being detached from the substratum, usually by invertebrate burrows (McLay and Hayward, 1987), age (specifically, by increased sedimentation and shading; Ghelardi, 1971), and wave forces (Dayton et al., 1984).
The objectives of this study were to quantify the variation in Hedophyllum sessile’s holdfast attachment properties and interpret the results in the context of H. sessile’s survival. Hedophyllum sessile is a dominant kelp species found in moderately to highly wave-exposed sites in the lower intertidal zone in the Northeastern Pacific Ocean (Abbott and Hollenberg, 1976).
Hedophyllum sessile sporophyte recruitment occurs in late winter through spring and
successful recruitment differs on different substrata types (Milligan, 1998). Juveniles are characterized by a holdfast, stipe, and blade (Widdowson, 1965). Holdfasts are made up of finger-like haptera that are attached to the substratum by microscopic tissue-extensions into crevices (Tovey and Moss, 1978). Juvenile mortality has been primarily attributed to herbivore damage by the dominant grazing chiton, Katharina tunicata (Duggins and Dethier, 1985). Adult mortality by holdfast dislodgment is primarily from winter storm damage (Dayton, 1975; Paine, 1980; Duggins and Dethier, 1985), and there is some evidence that K. tunicata will loosen adult holdfasts, making dislodgment more likely during winter storm swells (Markel and DeWreede, 1998). If thalli are cropped and tattered as a result of winter storms, the basal meristem and holdfast survive and new blade material will regenerate (Armstrong, 1987).
We investigated how attachment forces vary within Hedophyllum sessile at different developmental stages (juveniles vs. adults), wave-exposures, substratum types, and season. Since hydrodynamic force is bi-directional (shoreward and seaward) in the intertidal (Koehl, 1984), the holdfast mechanics in each of these directions were compared for adults. In addition, H. sessile’s holdfast removal force and strength were contrasted to those of other species found in similar environments to establish if H.
sessile displays either a strategy of strength or increased attached holdfast surface area to
resist wave forces. Causes for these patterns were not isolated in these experiments.
2. Materials and methods
2.1. Study area
Hedophyllum sessile was examined in moderately and extremely wave-exposed sites
Fig. 1. Map of study sites. All experiments were performed on Vancouver Island, British Columbia, Canada. Bamfield Marine Station is located on Trevor Channel, which opens to Barkley Sound and the Pacific Ocean. The approximate position of the LaPerouse Bank Buoy cited in Fig. 2 is represented by (B). Study sites were on either side of Prasiola Point (48849940N, 125810940W), which is marked by an arrow on the Trevor Channel / Bamfield map.
2.2. Biomechanical measurements: force to remove, strength, and thallus surface
area
Forces to remove Hedophyllum sessile were measured using a clamp and spring scale system. Juveniles were clamped just above the blade / stipe junction. Adults were clamped above the base of the blades. Juveniles were identified as individuals with a holdfast, stipe and blade, and adults were those individuals with holdfast and sessile blades. Two sizes of clamps (935 and 1538 cm) were constructed from flat plates of 0.63-cm thick exterior plywood lined with rubber (inner tubing, 2-mm thickness) to distribute force evenly and reduce slippage on the thallus. Each clamp thus consisted of two flat plates, each with drilled holes at either end to fit a 0.63-cm diameter nylon screw. Threaded knobs (Small Parts, Inc., Miami Lakes, FL; KK series phenolic knobs with brass inserts) on each screw tightened the plates around the thallus. Additional holes (two per plate) were drilled into the top of each plate, through which lines were threaded and used to attach the spring scales. Individual thalli were pulled parallel to the substratum, consistent with previous studies on algal attachment (Carrington, 1990; Shaughnessy et al., 1996). The place of breakage was noted for each sample, and then thalli were returned to the lab for morphological analyses.
Spring scales were modified from Bell and Denny (1994) by attaching a monofilament line to the clamp and using two different spring tensions: ‘heavy’ for adults (spring
21
extension constant51.990 N mm ) and ‘light’ for juveniles (spring extension
21
constant50.198 N mm ). Two scales were used in progression for the adults when the removal force exceeded the first scale. If removal caused both scales to reach maximum extensions, then the total force to remove was calculated to be the sum of the maximum extensions. This produced a conservative estimate of removal force and occurred in 20% of the samples. Scales were individually calibrated by hanging known masses (kg) and
22
multiplying by the gravitational acceleration, 9.81 m s , for a relationship between spring extension (mm) and force (N). The calibration curves were fit by linear regression analysis inSYSTAT5.03. Field data were spring extension (mm), which were converted to
force (N) by these calibration curves.
22
Strength (N m ) is defined as the force (N) required to break the attachment divided
2
by the cross-sectional area (m ) over which the failure occurs (Denny, 1988). Once the total force to remove was measured, each thallus was returned to the laboratory, blades were cut off at their bases, and all holdfast breaks directly scanned by placing the base of the holdfast on a flatbed scanner. These digitized images were then used to measure
2
the holdfast break area to nearest mm (Image 1.57, National Institute of Health). Juvenile holdfasts consisted of only a few haptera and could be freshly scanned. Adult holdfasts were made of more tightly compacted haptera and could not be scanned directly due to shadow effects. Thus, bases of break areas were photocopied and an outline drawn around the image on an acetate sheet, and these outlines were then scanned. In these cases, holdfast break area may have been slightly overestimated as
2
individual haptera were not measured. Areas were converted to m and strength
22
calculated as MN m .
2
manner as for the holdfasts. Total wetted area, m , was estimated by doubling the calculated blade area.
2.3. Substratum and exposure effects on juveniles
Experiments on the effects of substratum type and wave-exposure on juvenile attachment force were performed in June 1996. Juveniles were randomly chosen using a 10-m line with randomly marked distances placed horizontally through the middle of the
Hedophyllum sessile zone and the juvenile closest to each mark was chosen for
sampling. Randomly marked distances (cm) on the line were chosen using a random number generator. Juveniles were clamped and pulled shoreward in each site, and substratum type, force to remove, and break location were noted. Removal force and holdfast break area data for holdfast breaks were grouped into treatments of substratum type and site-exposure.
For removal force and holdfast area data, one-way ANOVA (in SYSTATversion 5.03)
was used to test for group differences, and a Tukey multiple comparisons test was used to detect which groups were significantly different. Data were log-transformed to attain normal distribution and homogeneity of variances. Mean holdfast strength was calcu-lated for each treatment.
To assess the percent substratum available in each site, substratum types were grouped within sites as articulated coralline algae (AC), crustose coralline algae (CC), non-calcareous red encrusting algae (RC), and bare rock (BR). Ten 0.2530.25 m plots were randomly placed by using a transect line with randomly marked distances (chosen using a random number generator) a minimum of 0.25 m. apart and alternating between the top-side and down-side of the line in the Hedophyllum sessile zone at each site. Percent cover of the substratum types was measured by point-intersect method (Dethier et al., 1993).
Substratum types varied in percent cover; to assess if recruitment differentially occurred between substratum types independently of the cover available by each substratum type, juvenile occurrences on the different substratum types during the biomechanical sampling were normalized by the percent cover of each substratum type. Differential recruitment (DR) was quantified by comparing the percent individuals sampled on each substratum to the percent cover of that substratum for recruitment, by calculating the following for each substratum type (S ):
21
DR5h(Ns?Nt )100j/ % S (1)
Where DR is differential recruitment, N is the number of individuals sampled ons
substratum type S, and N is the total number of individuals sampled on all substratat
types.
2.4. Site and seasonal wave-exposure effects on adults
To assess if wave-exposure acts to loosen holdfasts, thus reducing the total force to remove, effects of site and seasonal wave-exposure were measured on adult holdfast attachment. Adults were measured in wave-exposed and protected sites on Prasiola Point (48849940N, 125810940W) in July (pre-storm) and November (post-storm) 1996 (Fig. 2). At each site, horizontal transect lines were placed in the middle of the Hedophyllum
sessile zone and random points chosen along the lines, using same methods as used for
sampling juveniles. Adults closest to each point were chosen, and thalli were clamped and pulled. Pre-storm adults were pulled shoreward. In post-storm sampling, adults were pulled both shoreward and seaward at both wave-exposed and wave-protected sites to test if pull direction affects attachment.
Force to dislodge was measured and the break location was noted. Thalli were returned to the laboratory for thallus surface area and holdfast break area determinations. If the entire thallus was not dislodged (holdfast and blade material were left on the substratum), the remaining thallus was subsequently pulled to test if partial holdfast loss
would loosen the surviving portion. This occurred in less than 15% of individuals sampled.
Force and holdfast data for complete holdfast breaks in each pull-direction from the post-storm survey were log transformed to achieve normal distribution and homogeneity of variances. Data were analyzed by one-way ANOVA. Treatment groups were site exposures and pull directions. If no differences were detected in pull directions in each site, then all pull direction data within each site were pooled for further comparisons to pre-storm data.
To compare site and seasonal effects on holdfast attachment, force data from complete holdfast dislodgment were grouped into sites and seasons, log transformed to achieve normal distribution and homogeneity of variances, and tested for differences by a one-way ANOVA. The same procedure was used to test for seasonal effects on holdfast break area.
To further test how increased exposure affects variation of holdfast forces, cumulative probability distributions of normalized removal forces were constructed using methods of Gaylord et al. (1994), Denny (1995), and Friedland and Denny (1995). Each set of force data from exposed and sheltered sites in pre- and post-storm months was normalized by dividing the actual force to dislodge Hedophyllum sessile by the mean removal force for that set, yielding values for normalized breaking forces ( f ). f valuesn n
were ranked in ascending order; ties were assigned the same rank. The equation:
i
]]
P5 (2)
n11
was used to plot the cumulative probability (P) of a holdfast having less than a given f .n
The rank of the sample is represented by i, and n is the total number of samples. A generalized cumulative probability function
was fit to the observed probabilities from Eq. (2) using a nonlinear curve fitting routine in SigmaPlot version 2.0a. Confidence intervals (95%) were calculated for the cumulative distribution coefficients, a, b, and c in this curve-fitting routine. These intervals were used to determine if the estimated probability of holdfast dislodgment distributions for each site and season were different.
2.5. Relationship between holdfast and thallus size and attachment force
We predicted that larger holdfasts are more firmly attached and that larger holdfasts support more blade area. Pearson product–moment correlation analysis was used to test if there is a direct positive relationship between holdfast area and the force required to remove thalli. To test if larger holdfasts support more blade area, data from each sampling group were analyzed for a correlative relationship between holdfast area and blade area using Pearson product–moment correlation and regression analyses. Blade
2
2
per holdfast area (m ); averages and 95% confidence intervals for these ratios were calculated (Cochran, 1977).
3. Results
3.1. Substratum and exposure effects on juveniles
Relative to the percent cover of substrata types open for recruitment (Fig. 3), the number of juveniles found on each type were different (Table 1) and juveniles were most likely to be attached to articulated coralline algae. Only juveniles on articulated corallines (AC) in the exposed site and on articulated (AC) and crustose coralline algae (CC) in the protected site had sufficient sample size (n.2) and thus were the only sample groups used in analyses. Thus, juvenile attachment data were used to compare the difference between juvenile attachment on AC in exposed and protected sites, and to compare the attachment between AC and CC within the protected site.
For all substratum types and wave-exposure sites, juveniles primarily broke at the holdfast / substratum boundary (Fig. 4). All holdfasts completely broke off the sub-stratum, leaving no holdfast remnants; for this reason, holdfast break area is considered
Table 1
Differential juvenile recruitment by Hedophyllum sessile on different substrata types in May 1995, calculated
a
by Eq. (1)
Substratum type Differential recruitment Exposed site Protected site Articulated coralline algae 3.29 2.06 Crustose coralline algae 0.18 1.72 Non-calcareous encrusting red algae 0 0
Bare rock 0.37 0.38
Total sample size n536 n558
a
Differential recruitment is the percent of juveniles found on each substratum type normalized by the percent cover area of that substratum available for recruitment. Substratum bias for recruitment is demonstrated by differential recruitment values exceeding one.
equal to the holdfast area which was originally attached. Analysis of holdfast removal data (Fig. 5) showed that juveniles on articulated coralline algae (AC) in the exposed site required significantly more force to remove than those on AC in the protected site
Fig. 5. Forces (d) required to break juvenile holdfasts and the holdfast break areas (n) on different substrata and at two sites. See Fig. 4 for site and substratum abbreviations. Error bars are standard error. Asterisks above symbols represent groups that are significantly different (one-way ANOVA) from each other: *** P#0.001, and ** 0.001,P#0.01. Groups significantly different were: holdfast break area in protected site PRO,AC.
PRO,CC at 0.001,P#0.01; removal force from articulated coralline algae EXP,AC.PRO,AC at P#0.001; and removal force in the protected site PRO,AC.PRO,CC at 0.001,P#0.01.
(P,0.001, n523). Within the protected site, juveniles on AC required more force to remove than those on crustose coralline algae (CC; P50.003, n523) and had less juvenile holdfast area attached on CC than AC (P50.008, n523). Juveniles on CC, therefore, had weaker holdfast strength than juveniles on AC in the protected site (Table 2).
3.2. Patterns of adult attachment
Adult holdfast mechanics were quantified at pre-storm and post-storm dates. For both sampling dates and compared to the number of complete holdfast breaks, partial holdfast breaks occurred rarely in adult Hedophyllum sessile (Fig. 6). As with the juveniles, adult
H. sessile breaks primarily at the holdfast base (Fig. 7). However, upon removal, either small remnants of holdfast were left (holdfast-only break) or some of the coralline algae beneath the holdfast (holdfast / substratum break) was removed as well. Since both of these break-types resulted in complete thallus dislodgment, both of these types of breaks were considered holdfast breaks and pooled for statistical analyses.
Table 2
22 a
Hedophyllum sessile holdfast attachment strengths (MN m 6S.E.) of sporophyte juveniles and adults 22
Sample group Holdfast attachment strength (MN m 6S.E.) n Juveniles
Exposed, AC 0.02360.002 32
Protected, AC 0.01260.002 31
Protected, CC 0.00560.002 23
Adults
Exposed, pre-storm 0.07060.011 17
Exposed, post-storm 0.07860.012 17
Protected, pre-storm 0.07360.015 17
Protected, post-storm 0.04560.005 17
a
AC represents juveniles on articulated coralline algae and CC represents juveniles on crustose coralline algae; n is the sample size within each sample group.
different in removal force (P50.279, n511) or holdfast break area (P50.797, n511). Shoreward and seaward pulls were pooled and used to compare force to break and break area between sites and seasons.
Fig. 7. Percent occurrence of different break locations for adults at two sites in pre- and post-storm sampling dates. Data include only complete holdfast breaks from Fig. 6 and do not include partial holdfast breaks. Breaks are abbreviated as HO, holdfast only (total holdfast separates from the substratum); H / S, holdfast and substratum (total holdfast detaches with the underlying substratum); TTL HO, holdfast only1holdfast / substratum; BL, blade (holdfast does not detach but the blade breaks). Samples from the exposed site are represented by solid bars and those from the protected site are represented by open bars.
The force (pooled mean5100.69 N66.79 S.E.) required to remove H. sessile adults was consistent over sites and seasons (one-way ANOVA, P50.355, n513; Fig. 8).
2
Holdfast break area (pooled mean50.0019 m 60.0001 S.E.) was also consistent over sites and seasons, with the exception that the protected, post-storm group was significantly larger than the exposed, pre-storm group (one-way ANOVA, P50.025,
n513). Thus, the protected, post-storm group had a weaker holdfast attachment strength than other experimental groups (Table 2). There was no direct relationship between holdfast break area and the force to remove in the pre-storm sampling, but a significant positive relationship existed in the storm sampling at both sites (exposed post-storm: R50.612, P50.009, n517; protected post-storm: R50.563, P50.019, n517). Larger holdfasts supported more thallus area (Fig. 9), and holdfasts at both sites support significantly less blade area in the post-storm sampling than in the pre-storm sampling date (Table 3).
Fig. 8. Forces (d) required to break adult holdfasts and the holdfast break areas (n) at two sites in pre- and post-storm sampling dates. Sites are labeled as EXP, wave-exposed site; PRO, wave-protected site. Error bars are standard error. Asterisks above symbols represent the sampling dates and sites (for holdfast break area, pre-storm EXP,post-storm PRO) that are significantly different, 0.01,P#0.05 (one-way ANOVA).
detached, there was preliminary evidence that the remaining holdfast on the rock was attached less firmly. However, sample sizes were too small for statistical analyses; therefore, it can not be concluded that partial removal affects the integrity of the remaining holdfast.
Even though analysis of attachment properties using one-way ANOVA showed no differences in average attachment forces among sites and seasons (Fig. 8), there is a seasonal effect on the distribution of attachment forces within the sampled populations. Despite no change in holdfast size, within each site, there is a shift to more resistant holdfasts as a result of increased seasonal wave-exposure (Fig. 10). For simplicity, 95% confidence intervals (Table 4) are not shown around the estimated distributions in Fig. 10; however, above the normalized force of 1.5, there is a seasonal effect at each site (P,0.05) towards more resistant holdfasts where the post-storm population had a lower probability of removal than the pre-storm population.
4. Discussion
4.1. Juvenile attachment
Fig. 9. Relationship between holdfast and wetted blade surface area. Each data point represents an individual thallus sampled for holdfast attachment at exposed (dm) and protected (sn) sites in pre-storm (ds) and post-storm (mn) dates. Solid (wave-exposed site) and dashed (wave-protected site) lines are the regression lines for the significant, direct relationships between holdfast area and the wetted blade surface area that it supports (refer to Table 3 for the correlation statistics and sample sizes).
interactions are potentially important factors for determining juvenile survivorship and abundance. Hedophyllum sessile juveniles occur more often on articulated coralline algae (AC) than on other available substrata and this difference is more distinct in the exposed site, as evidenced by differential recruitment estimates.
An organism’s attachment force to the substratum can be increased by two mechanisms: by an increase in strength and / or by an increase in attachment surface area.
Table 3
a
Analysis of wetted blade surface area per holdfast area
Sample group Wetted blade surface area / n R P value holdfast area ratio (695% CI)
Exposed, pre-storm 157.10666.49 15 0.67 0.007
Protected, pre-storm 167.90654.41 18 0.84 ,0.001
Exposed, post-storm 62.27615.19 21 0.65 0.001
Protected, post-storm 76.98613.21 33 0.75 ,0.001
a 2 2
Fig. 10. Cumulative functions for probability of holdfast dislodgment. Each data point represents the observed normalized attachment forces ( f ) at exposed (n dm) and protected (sn) sites in pre-storm (ds) and post-storm (mn) dates. Solid (wave-exposed site) and dashed (wave-protected site) lines are the calculated probability distributions from Eq. (3) for each sample date as noted. Equation coefficients695% CI are given in Table 4.
Hedophyllum sessile juveniles had a significantly greater attachment force on AC algae
in the exposed site than in the protected site. On AC, juveniles in the exposed site had less holdfast area and consequently greater strength than the juveniles in the protected site. Thus, these results demonstrate that H. sessile juveniles have an increased total
Table 4
Coefficient values (695% CI) for the cumulative distributions for the probability of breaking Hedophyllum sessile holdfast attachments in exposed (E) and protected (P) sites in pre- and post-storm months (see Eq. (3) for description of the coefficients)
attachment force to the substratum by increased attachment strength rather than by increased holdfast area.
These observed differences in juvenile attachment forces and strengths between sites and substrata types are likely a reflection of physical mechanisms that begin acting at microscopic stages of development. Rougher surfaces have been shown to provide optimal settlement and growth environments for some species, such as Enteromorpha sp. (Christie and Shaw, 1968) and Sargassum muticum (Norton, 1983) by offering refuge from shear forces and more surface rugosity for holdfast development and attachment. Holdfasts (in Laminariales) attach by forming microscopic rhizoids from their haptera that fill in crevices on the substratum (Tovey and Moss, 1978). Thus, any greater surface roughness and microsurfaces of AC in comparison to other substrata may provide more interstitial areas into which rhizoids may penetrate and thus provide extra attachment strength (force per unit area) and total holdfast force.
Previous studies on Hedophyllum sessile recruitment have focused on effects of grazers, specifically the chiton Katharina tunicata, on juvenile distributions and densities (Duggins and Dethier, 1985; Markel and DeWreede, 1998). Articulated coralline algae have upright thallus parts and they form a low (,10 cm) turf. H. sessile holdfasts are attached at the base of the turf. In this study, we can not exclude grazer influence as a potential mechanism of juvenile H. sessile distributions since AC may provide a refuge from K. tunicata (Markel and DeWreede, 1998).
Results here suggest that another possible driving mechanism for this distribution pattern is hydrodynamic stress, because juveniles increase holdfast strength and thus total attachment force in AC turfs and in correlation with increased wave exposure. Significantly greater attachment force and greater strength of Hedophyllum sessile juveniles on AC algae in the exposed site than in the protected site indicate a biomechanical response to wave-exposure, whereby more hydrodynamic stress is reflected in firmer attachment properties. The lack of juveniles on crustose coralline algae (CC) in the exposed environment, but their presence in the more protected site, could be explained by the fact that haptera are not attached firmly on CC, which may result in greater dislodgment risks in the exposed site than in the protected site. Given these findings, the potential for interactive effects between grazing, substratum type, and juvenile holdfast attachment properties on resultant recruitment warrant further in-vestigations.
4.2. Adult attachment
In comparison to other algal taxa for which field attachment forces have been measured, Hedophyllum sessile has a weak attachment strength, but extremely high attachment force attained by large surface area. This is in contrast to patterns observed in juvenile populations, where greater attachment force is achieved by increased strength. More than 85% of the adult H. sessile holdfasts detached completely. The holdfast
22
attachment strength of H. sessile is 0.07 MN m whereas, for example, the stipe / blade
22
attachment strength for Mazzaella linearis and M. splendens is 8–9 MN m
(Shaug-22
holdfast dislodgment; 100.68 N) is substantially higher than other taxa which have been documented, with the exception of kelp species such as Egregia menziesii (stipe break589 N; Friedland and Denny, 1995).
Gaylord et al. (1994) predicted that large holdfasts are selected against, most likely because they support more biomass. In this study, selection in Hedophyllum sessile was against large but loose holdfasts and thallus tattering was a common occurrence. Within each site, there was no difference between sampling dates of average holdfast sizes but the mean blade surface area was lower in the post-storm sampling, which is likely a result of thallus tattering, not of dislodgment. Initially, it was predicted that holdfast size would be directly correlated to its removal force. Likewise, Norton et al. (1982) found that, in Macrocystis pyrifera, small holdfasts less than 2 years old were less firmly attached than those older with larger holdfasts. However, our results illustrate that the relationship between size and attachment force is not always directly correlated, and that a correlation is seasonally dependent as well as size (and age)-dependent. Larger holdfast surface area did not correlate to higher attachment force in the pre-storm sampling date, suggesting that holdfasts loosen as they become larger and that, prior to winter storms, large holdfasts are disproportionally weak. In the post-storm sampling, holdfast areas were significantly correlated to the force required to dislodge them, suggesting that the large but loose holdfasts were removed from the population. This loss of large but loose holdfasts resulted in the probability of holdfast dislodgment distributions within the populations shifting towards more resistant holdfasts in the post-storm sampling.
Biomass loss by thallus tattering and resistance to detachment by holdfast attachment force varied as a result of increased storm-swells and hence wave-exposure. Variations in the biomass relative to holdfast attachment force will place different detachment (and mortality) risks on adult Hedophyllum sessile during hydrodynamic stress. Our results indicate that biomechanical parameters (such as holdfast attachment force and drag forces based on wetted area) measured for populations can not be assumed to be constant. These biomechanical properties will be exposure-specific (e.g. with site or seasonal exposure) as demonstrated in this study. Future studies should attempt to document the seasonal variation of biomechanical attributes in macroalgal field populations, especially when these parameters are used to interpret demographic rates such as survival across a range of sites.
Extensive field sampling to establish how attachment properties vary within popula-tions is labor intensive and destructive. The ability to predict the extent of holdfast-loosening events would be an important step towards predicting seasonal dynamics of attachment properties. However, mechanisms creating large, loose holdfasts have not been well quantified. Loosening agents such invertebrate burrowing activities and substratum failure may be important factors contributing to holdfast failure. Even though interactions between a kelp holdfast’s invertebrate community and the growth of the holdfast have been reported (Ojeda and Santelices, 1984; McLay and Hayward, 1987), the subsequent effects on attachment forces and survival of the kelp have been rarely documented (Tegner et al., 1995).
coralline algae were removed with the Hedophyllum sessile haptera. Assuming that larger holdfasts are older, coralline algae under larger holdfasts may be degraded sufficiently to increase the chance of holdfast dislodgment by substratum failure.
4.3. Ecological implications: recovery by juveniles after adult removal
Studies on thallus dislodgment have demonstrated that removal occurs when hydro-dynamically-induced force exceeds attachment forces (Gaylord et al., 1994; Friedland and Denny, 1995). Biomass loss in Hedophyllum sessile would ultimately reduce stress on the holdfasts. If dislodgment occurs, H. sessile holdfasts are the most likely location for failure and will most frequently be completely removed with no chance of regeneration. When adults are removed, part of the failure is at the crustose substratum, and consequently a patch of bare rock is opened for colonization. It is rare to find juveniles on bare rock in these moderately to highly wave-exposed sites (Milligan, 1998); recolonization by H. sessile may be most successful once coralline algae have re-established from the surrounding area. If articulated coralline algae colonize the open area, then this will enhance H. sessile recruitment and offer a more secure attachment surface than crustose coralline algae.
5. Conclusion
This is the first report of the role of physical forces on Hedophyllum sessile juvenile distribution patterns; juveniles display increased strength in a more wave-exposed site in comparison to a less wave-exposed site and on articulated coralline algae in comparison to strength on crustose coralline algae. The data presented here also demonstrate that results from holdfast attachment field studies in one season cannot be extrapolated to another due to a complex set of dynamics. For adult Hedophyllum sessile, seasonal increases in storm swells resulted in lower blade surface area by thallus tattering and more resistant holdfasts by selection against large, loose holdfasts. Comparisons of distributions of probability of dislodgment illustrated a shift towards more resistant holdfasts after a series of storms. This is the first documentation of seasonal shifts in macroalgal attachment properties. Since an alga’s attachment is essential to its survival, this research provides motivation for future research on seasonal effects on algal attachment, the extent to which results from one season can be applied to other seasons, the extent to which these variations affect survival, and the driving factors for these patterns.
Acknowledgements
Markel, Dr Ricardo Scrosati, Dr Andrea Sussman, and Tania Thenu. This work was greatly improved by discussions with Drs Emily Bell, Bryan Gaylord, John Gosline, Catriona Hurd, Allen Milligan, Frank Shaughnessy, and Craig Stevens and review by Drs Mark Denny, John Gosline, Paul G. Harrison, Roy Turkington, and comments by anonymous reviewers. Financial support came from the Department of Botany, University of British Columbia, NSERC (grant [ 5-89872 to R.E. DeWreede), and Bamfield Marine Station (summer graduate student research support-grant to K. Milligan). [SS]
References
Abbott, I.A., Hollenberg, G.S., 1976. Marine Algae of California. Stanford University Press, Stanford, CA. Armstrong, S.L., 1987. Mechanical properties of the tissues of the brown alga Hedophyllum sessile (C.Ag.)
Setchell: variability with habitat. J. Exp. Mar. Biol. Ecol. 114, 143–151.
Bell, E.C., Denny, M.W., 1994. Quantifying ‘wave-exposure’: a simple device for recording maximum velocity and results of its use at several field sites. J. Exp. Mar. Biol. Ecol. 181, 9–29.
Biedka, R.F., Gosline, J.M., DeWreede, R.E., 1987. Biomechanical analysis of wave-swept mortality in the marine alga Pterygophora californica. Mar. Ecol. Prog. Ser. 36, 163–170.
Carrington, E., 1990. Drag and dislodgment of an intertidal macroalga: consequences of morphological variation in Mastocarpus papillatus. J. Exp. Mar. Biol. Ecol. 139, 185–200.
Christie, A.O., Shaw, M., 1968. Settlement experiments with zoospores of Enteromorpha intestinalis (L.) Link. Br. Phycol. Bull. 3, 529–534.
Cochran, W.G., 1977. Sampling Techniques, 3rd Edition. Wiley, New York.
Dayton, P.K., 1975. Experimental evaluation of ecological dominance in a rocky intertidal algal community. Ecol. Monogr. 45, 137–159.
Dayton, P.K., Currie, V., Gerrodette, T., Keller, B.D., Rosenthal, R., Ven Tresca, D., 1984. Patch dynamics and stability of some California communities. Ecol. Monogr. 54, 253–289.
Denny, M.W., 1988. Biology and the Mechanics of the Wave-swept Environment. Princeton University Press, Princeton, NJ.
Denny, M.W., 1995. Predicting physical disturbance: mechanistic approaches to the study of survivorship on wave-swept shores. Ecol. Monogr. 65, 371–418.
Denny, M., Brown, V., Carrington, E., Kraemer, G., Miller, A., 1989. Fracture mechanics and the survival of wave-swept macroalgae. J. Exp. Mar. Biol. Ecol. 127, 211–228.
Dethier, M.N., Graham, E.S., Cohen, S., Tear, L.M., 1993. Visual versus random-point percent cover estimations: ‘objective’ is not always better. Mar. Ecol. Prog. Ser. 96, 93–100.
DeWreede, R.E., Ewanchuk, P., Shaugnessy, F., 1992. Wounding, healing and survivorship in three kelp species. Mar. Ecol. Prog. Ser. 82, 259–266.
Dudgeon, S.R., Johnson, A.S., 1992. Thick vs. thin: thallus morphology and tissue mechanics influence differential drag and dislodgment of two co-dominant seaweeds. J. Exp. Mar. Biol. Ecol. 165, 23–43. Duggins, D.O., Dethier, M.N., 1985. Experimental studies of herbivory and algal competition in a low
intertidal habitat. Oecologia 67, 183–191.
Friedland, M.T., Denny, M.W., 1995. Surviving hydrodynamic forces in a wave-swept environment: consequences of morphology in the feather boa kelp, Egregia menziesii (Turner). J. Exp. Mar. Biol. Ecol. 190, 109–133.
Gaylord, B., Blanchette, C.A., Denny, M.W., 1994. Mechanical consequences of size in wave-swept algae. Ecol. Monogr. 64, 287–313.
Ghelardi, R.J., 1971. ‘Species’ structure of the animal community that lives in Macrocystis pyrifera holdfasts. Nova Hedwigia 32, 381–419.
Koehl, M.A.R., 1984. How do benthic organisms withstand moving water? Am. Zool. 24, 57–70. Koehl, M.A.R., Wainwright, S.A., 1977. Mechanical adaptations of a giant kelp. Limnol. Oceanogr. 22,
1067–1071.
McLay, C.L., Hayward, T.L., 1987. Population structure and use of Durvillaea antarctica holdfasts by the intertidal spider Desis marina (Araneae: Desidae). NZ J. Zool. 14, 29–42.
Markel, R.W., DeWreede, R.E., 1998. Mechanisms underlying the effect of the chiton Katharina tunicata (Wood) on the kelp Hedophyllum sessile (C. Agardh) Setchell: size escape and indirect effects. Mar. Ecol. Prog. Ser. 166, 151–161.
Milligan, K.L.D., 1998. Effects of wave-exposure on an intertidal kelp species Hedophyllum sessile (C.Agardh) Setchell: demographics and biomechanics. PhD thesis, Department of Botany, University of British Columbia, Vancouver, British Columbia, Canada.
Norton, T.A., 1983. The resistance to dislodgment of Sargassum muticum germlings under defined hydrodynamic conditions. J. Mar. Biol. Assoc. 63, 181–193.
Norton, T.A., Mathieson, A.C., Neushul, M., 1982. A review of some aspects of form and function in seaweeds. Bot. Mar. 25, 501–510.
Ojeda, F.P., Santelices, B., 1984. Invertebrate communities in holdfasts of the kelp Macrocystis pyrifera from southern Chile. Mar. Ecol. Prog. Ser. 16, 65–73.
Paine, R.T., 1980. Food webs: linkage, interaction, strength, and community infrastructure. J. Anim. Ecol. 49, 667–685.
Shaughnessy, F.J., DeWreede, R.E., Bell, E.C., 1996. Consequences of morphology and tissue strength to blade survivorship of two closely related Rhodophyta species. Mar. Ecol. Prog. Ser. 136, 257–266.
Tegner, M.J., Dayton, P.K., Edwards, P.B., Riser, K.L., 1995. Sea urchin cavitation of giant kelp (Macrocystis pyrifera C. Agardh) holdfasts and its effects on kelp mortality across a large California forest. J. Exp. Mar. Biol. Ecol. 191, 83–99.
Tovey, D.J., Moss, B.L., 1978. Attachment of the haptera of Laminaria digitata (Huds.) Lamour. Phycologia 17, 17–22.
