provide refuge from predation, alleviate physical or physiological stress and control transport of solutes and particles in the benthic environment (Gutiérrez et al. 2003).
In field studies, Aguilera et al. (2017) examined the attachment strength of the mussels Perumytilus purpuratus, growing on Pacific hard‐ and soft‐bottom shores in Chile, and M. edulis, growing on an Atlantic rocky shore in France and a sedimentary shore in the North Sea (Germany). They found that the attachment strength of P. purpuratus on hard bottoms was substantially higher than that on soft bottoms even though mussels produced more byssus in the latter habitat. In contrast, M. edulis showed only a slightly reduced attachment strength on soft compared to hard bottoms. In laboratory experiments, the authors examined the mussel substratum selectivity of both bivalve species from soft and hard bottoms by offering living versus dead, barnacle‐fouled versus unfouled and firmly attached versus loose conspecifics. In these experiments, P. purpuratus from both habitats showed a significant preference for living conspecifics, while M. edulis had no preference for particular conspecifics – except those that preferred fouled over clean mussels from soft‐bottom habitats. In summary, their results confirmed that marine mussels show active substratum choice for byssus attachment, which depends on mussel species and habitat type. Also, their results suggest that mussels are adapted to a particular habitat type, with P. purpuratus showing lower adaptation to soft‐bottom areas and M. edulis showing equal preference for both substrate types. In a previous study, Thiel & Ullrich (2002) described the fauna associated with hard‐bottom mussel (P. purpuratus) beds at eight sites along the Pacific coast of Chile, with the aim of elucidating the functional role of mussel beds in hard‐ and soft‐bottom environments. At all sites, the associated fauna was dominated by suspension feeding organisms (cirripeds, spionid and sabellid polychaetes), followed by grazing crustaceans and gastropods. Predators and scavengers also reached high abundances, while deposit‐ and detritus‐feeding organisms were of minor importance. The majority of organisms associated with these hard‐bottom mussel beds feed on resources obtained from the water column, or grow on the mussels rather than on materials (faeces, pseudofaeces) deposited by the mussels. This is in contrast to the fauna associated with mussel beds on soft bottoms, which comprise many species feeding on material accumulated by mussels and deposited within the mussel bed, indicating that mussels on hard bottoms primarily provide substratum for associated fauna, while mussels on soft bottoms provide both substratum and food resources.
Growth and mortality rates and inducible defence characters on medium‐sized M. edulis (18–22 mm shell length) exposed to shore crab (Carcinus maenas) predation were examined on three different substrate types in combined field and laboratory experiments (Frandsen & Dolmer 2002). The substrate types used were: a smooth substrate, classified as simple, that structurally resembles sand; unbroken shells, classified as complex; and live M. edulis, classified as complex. High complexity and heterogeneity of a substrate is believed to reduce predation pressure by increasing the number of spatial refuges (references in Frandsen & Dolmer 2002). The experiments showed that crab predation was significantly higher (one‐way ANOVA, P < 0.001) on the smooth substrate compared to the two more complex substrates, with no significant difference in predation between the complex types. However, increased intraspecific competition for food on the complex substrates resulted in significantly lower growth rates of the mussels. Inducible defence characters were also influenced by substrate type. Mussels were more affected by predators on the structurally simple substrate, where they developed thicker shells and a significantly (P < 0.01) larger posterior adductor muscle, both of which are defence responses that cause predators to take much longer to open the mussels (Freeman 2007). Finally, interspecific substrate preferences have been described in Gilg et al. (2010, ch. 5) and Katolikova et al. (2016, ch. 9).
Disturbance
The lower limit in mussel beds is typically controlled by the interaction between recruitment and sea star predation, and the upper limit by desiccation (Robles & Desharnais 2002). But within the mussel bed itself, the ability of mussels to dominate space is limited primarily by the dislodgement of individuals by waves (see earlier). Dislodgment opens patches (gaps) of bare substratum in the bed, temporarily providing space for fugitive species, which are eventually snuffed out by reinvasion of the bed (Denny & Gaylord 2010). This process, referred to as ‘disturbance’, has been defined as ‘the displacement, damage or death of organisms caused by an external physical force or condition or incidentally by a biological entity’ (Sousa 2007, p. 186). Most of the information on the effects of, and recovery from, disturbance comes from studies of Mytilus beds and their associated flora and fauna on exposed shore sites on the Pacific and NE Atlantic coasts of North America (Seed & Suchanek 1992; Svane & Ompi 1993; Wootton 1993; Beukema & Cadee 1996; Carroll & Highsmith 1996; Hunt & Scheibling 1998, 2001; Bertness et al. 2002; Guichard et al. 2003; Sousa 2007, 2012; Calcagno et al. 2012).
Apart from wave action, common physical agents that initiate gap formation include impact or abrasion by waveborne objects such as cobbles, logs or ice; extremes of air or water temperatures; desiccation; fouling by brown algae and barnacles; hummocking; abrasion by suspended sand; and burial under deposited sand (Sousa 2007). Wave and log damage occurs mostly during the winter and is typically responsible for removing 1–5% of M. californianus cover per month on exposed shores (Paine & Levin 1981). The initial size of disturbance gaps can range from single mussel size to areas as large as 60 m2. Subsequent enlargement of the gap (as much as 5000%) may occur, especially during winter months, primarily due to weaker byssal thread attachments (Witman & Suchanek 1984). While some mussel beds remain uniform and flat, others form elevated hummocks consisting of small groups of 10–20 mussels, which become detached from the rock and are forced upward above the bed surface. Mussels (Brachidontes rodriguezii) in hummocks show lower attachment strength than those in the single‐layered matrix (Gutiérrez et al. 2015). Accordingly, wave conditions associated with the passage of cold fronts (i.e. transition zones from warm air to cold air accompanied by moderate to strong winds and wave action, with seven‐day average recurrence times based on historical weather data) caused detectable mussel dislodgment in a high proportion of hummocks but have virtually no impact on single‐layered areas. Biological disturbances that disrupt the matrix of Mytilus beds are predators such as crabs and sea stars and epizoism by algal fronds; these usually occur on the scale of individual or a few mussels, but the sea star Picaster forms larger gaps in M. californianus beds (Seed & Suchanek 1992).
Patch structure and dynamics will probably differ among habitat types because of environmental differences influencing the intensity and outcome of biological interactions. Hunt & Scheibling (2001) compared the dynamics of natural and experimentally constructed mussel patches (M. trossulus and M. edulis) in two intertidal habitats, tidepools and emergent rock, over a time series (5, 10, 15 months) and among seasons (three successive five‐month intervals). In tidepools, mussels naturally occurred in small patches (median <25 cm2), while mussels on emergent rock formed extensive beds with decimetre‐scale gaps, but these beds started as small patches following disturbance such as ice scour. For their experimental patches (15 cm2) in both habitats, the authors assessed the relative importance of physical (wave disturbance) and biological (predation, growth, recruitment and immigration) processes in determining patch size and structure. Individual experimental and natural patches varied greatly in size over time, but mean patch area remained relatively constant. Mean size of individuals in experimental patches decreased due to the loss of larger mussels, while numbers increased due to recruitment. Wave disturbance appeared to be more important than predation in determining patch structure and dynamics, although losses due to either process did not differ consistently between habitats over time. Growth rates were low (≤0.4mm per month), but were greater in tidepools than on emergent rock, whereas recruitment and immigration rates generally did not differ between habitats. Although each process contributed to changes in patch size and structure, overall they did not result in a marked divergence in mean patch area or biomass between tidepools and emergent rock over the 15‐month experiment. This study highlights how integrative approaches,