Four primary estuarine/marine ecotopes recognized in the North Atlantic continental littoral and continental terrestrial margins give rise to the following ecotypes: wrack generalists (=beach-hoppers), psammophilic talitrids (=sandburrowing-hoppers), palustral talitrids (=salt marsh-hoppers), and xylophagous talitrids (=driftwood hoppers). On the European continent, there are freshwater riverine and lacustrine talitrids. In addition, there are a few terrestrial ecotypes in the Northeast Atlantic Islands: rainforest leaflitter talitrids and one troglobiont (=cave-hopper). Wrack generalist species are phenotypically plastic and can live in one or more secondary ecotopes. It is hypothesized herein that switching on/off appropriate genes by cellular mechanisms (epigenesis) occurs during the microevolution of Talitridae. The generalist/specialist continuum concept supports our understanding of both the ecology and microevolution of talitrids. Microevolutionary characteristics of wrack generalists are that they exhibit the most phenotypic variability and occurrences of epigenesis, have the most extensive zoogeographic range, and have the lowest speciation potential and endemism rate. Examples where epigenesis may be part of the microevolutionary process include low/high salinity and hypogean/epigean combined switches giving rise to sibling (sister) species pairs. Current views of the phylogeny of the Talitridae based on either morphological characters used in taxonomy or molecular genetic methods are still under development. Molecular genetic methods show promise of providing a scientifically reproducible phylogeny and temporal history of talitrids (macroevolution), but insufficient coverage of genera within talitrids and of related groups is available to do so yet.

Generalist/specialist continuum, micro-/macro-evolution, primary and secondary talitrid ecotypes, sibling species pairs, Talitridae

The Talitridae (talitrids) are a family within the Amphipoda, which in 2018 consisted of 80 genera and 512 species worldwide (Tykarska et al. 2019), with new genera and species being added annually. Many of the new talitrid species described do not contain adequate ecological descriptions of the type locality (Lowry and Myers 2019).

Talitrids are found in many ecotopes (=ecological habitats or niches) from marine and estuarine to freshwater to terrestrial (Hurley 1968; Friend and Richardson 1986; Wildish 1988). A current summary of the established ecotopes occupied by talitrids on the shores of the Atlantic Ocean is shown in Table 1. This is a preliminary list because only a few talitrids, mostly in the temperate North Atlantic, have been adequately studied from an ecological perspective. As further ecological studies, notably on southern hemisphere terrestrial species, are completed, further ecotopes will be added to a world list expanded from Table 1.

Evolutionary studies of talitrids include changes that occur at very different temporal and spatial scales. Both scales are continuous, making it difficult to characterize the particular evolutionary changes of concern. Here I consider microevolution to result from a limited number of environmental factors and genetic or epigenetic changes in one or a few species. By contrast, macroevolution involves all species within the family Talitridae from their origin millions of years ago from an amphipod stem group to their radiation to most parts of the world today.

This presentation is an update and expansion of a review dealing only with xylophagous talitrids (Wildish 2017). The expansion includes new ecological discoveries and microevolutionary findings concerning Talitridae in all ecotopes of the North Atlantic coastal region. This includes the Gulf of Mexico and the Mediterranean Sea, Atlantic islands, and the terrestrial ecotopes of the Northeast Atlantic islands (Azores, Madeira, and Canaries Archipelagos).

In ecology, we require an accurate descriptive nomenclature for each talitrid ecotype (a talitrid species occupying and adapted to a particular ecotope) recognized so that it is possible to refer each species to an ecotope within a given ecosystem (Table 1). For this reason, the term “semi-terrestrial” as occasionally used in the literature for marine intertidal talitrids (e.g., Hupalo and Grabowski 2018) is inadmissible in ecology because there is no ecosystem to which it is referable. Thus, wrack generalists (as here defined) are never “semi-terrestrial” and are more accurately described in an ecological sense as ‘amphibious’ because they live in both damp air in the intertidal zone and sometimes in seawater (at high tide and occasionally during passive seawater dispersal). Table 1 considers only primary ecotypes and represents a work in progress, as it is expected that many more ecotopes and ecotypes remain to be discovered and described from other parts of the world. Primary ecotopes are ones in which a given species is usually found and to which it has evolved characteristic adaptations. It is contrasted with one or more secondary ecotopes, in each of which the same species is less commonly found and lacks some of the characteristic adaptations of the primary ecotype (Wildish and Robinson 2016a).

The most studied ocean system concerning talitrid ecology is the coastal North Atlantic, inclusive of the Gulf of Mexico and the Mediterranean Sea, with a species and ecotype list given by Wildish and Chang (2017). This list is summarized in Table 2. It does not include the northeast Atlantic Island talitrids; they are dealt with separately below. There is disagreement regarding the zoogeography of Orchestia gammarellus (Pallas, 1766) in the northwest Atlantic (Canada and Iceland), with Myers and Lowry (2020) suggesting that it is a new species despite the molecular genetic evidence of Henzler and Ingolfsson (2008) that it is the same species on both sides of the North Atlantic. Wildish and McDonald (2023) have proposed a natural, passive dispersal hypothesis to account for the presence of O. gammarellus in post-glacial Iceland and Canada. The other amphi-Atlantic species of Table 2 are Platorchestia exter Myers & Lowry, 2023, and P. oliveirae Myers & Lowry, 2023. These two species are the dominant wrack generalists of the coastal northwest Atlantic with a zoogeographic range from Brazil to Newfoundland. Platorchestia exter occupies the northerly part and P. oliveirae the more southerly part of this range, with the separation occurring somewhere in, or north of, Florida. It has been hypothesized (Wildish and Chang 2017) that both species reached the northeast Atlantic coast recently either by natural, passive rafting carried by Gulf Stream currents or by synanthropy. Preliminary molecular genetic evidence that the European Platorchestia are the same species as those from North America is available (Radulovici 2012; Hupalo and Grabowski 2018).

It was discovered by field observations in 2015 that P. exter could occupy a secondary ecotope: stranded driftwood (Wildish and Robinson 2016a) when specimens were found occupying this habitat in a driftwood depository in the Bay of Fundy. This study also confirmed that P. exter could be cultured solely on driftwood as food for at least six months. Driftwood acclimation, which involves ingesting rotting driftwood, resulted in reduced growth and shorter adult total body lengths (Table 3). In the control culture, treatment C of Table 3, total body length is significantly greater than in treatment A by t-test (t = 11.225, df = 167, p < 0.001), and C is also greater than B (t = -2.952, df = 73, p < 0.005). It was experimentally shown that slower growth rates were linked to lower standard metabolic rates (Wildish and Robinson 2018), suggesting that aged driftwood was more difficult to digest than aged macroalgal wrack. Experimental demonstration that P. exter can readily reverse the acclimation process from driftwood to wrack or vice versa has been obtained (Wildish and Robinson 2018). The mechanisms that produce dwarf morphs within P. exter because of a diet switch to driftwood remain unknown but could involve some form of phenotypic plasticity (see below).

Dwarf secondary ecotypic morphs have also been recognized in two other wrack generalists. The eulittoral wrack generalist Orchestia mediterranea A. Costa 1857 was found in a floating piece of driftwood that beached in the tidal Swale, Kent, England, containing 16 specimens (Wildish 2018 and references therein). The dwarf specimens of O. mediterranea were originally described as a new xylophagous specialist species, but the genetic data showed this was not the case (Wildish 2018). It is possible that O. gammarellus can also exist in a secondary driftwood ecotope (Wildish 2017). Thus, at least three species of secondary ecotype dwarf morphs associated with xylophagy have been identified and the primary ecotype of each is a wrack generalist, suggesting that this may be a general feature of this ecotype.

Orchestia gammarellus is also reported to live in European salt marshes (Sprung and Machado 2000; Dias and Sprung 2003; Schrama et al. 2015) rather than in the supralittoral drift line of beaches (primary ecotope), and this may be a further example of a secondary palustral ecotype. As far as I am aware, the morphological, physiological, and behavioral changes associated with a palustral ecotope for O. gammarellus have not been investigated. However, a change of diet from decaying macroalgae to decaying lignivorous salt marsh grasses is involved.

A generalist/specialist continuum within Talitridae was introduced by Wildish and Radulovici (2019) to help understand the ecological and evolutionary characteristics of this family. The generalist group of species are those capable of living in more than one ecotope and consuming a varied diet range. They contrast with specialist groups that live in a single ecotope and consume a singular diet.

Specific examples of primary ecotypes currently known from the North Atlantic coastal region are listed below in the same continuum order (generalist to specialist) shown in Table 4:

• # 1. Wrack generalists include P. exter, P. oliveirae, and O. gammarellus, which occur on both northeast and northwest Atlantic coasts. Wrack generalists limited to the northwest Atlantic coast are Mexorchestia carpenteri carpenteri, M. c. raduloviciae, Tethorchestia antillensis, and Platorchestia sp. (identified genetically by Radulovici 2012). Limited to the northeast Atlantic coast are Orchestia mediterranea, O. aestuarensis, O. montagui, O. xylino, and Speziorchestia stephensi.
• # 2. Up to 7 genetically defined palustral species currently included within Speziorchestia grillus are confined to the northwest Atlantic coast.
• # 3. Different species of psammophilic talitrids occur on each side of the Atlantic: northwest: Americorchestia longicornis, A. megalophthalma, A. heardi, A. salomani, and A. barbarae; on the northeast: Talitrus saltator, Deshayesorchestia deshayesii, Sardorchestia pelecaniformis, Britorchestia brito, B. ugolini, and Africorchestia spinifera.
• # 4. Five species of xylophagous talitrids occur on the northeast Atlantic coast only: Macarorchestia microphthalma, M. remyi, M. pavesiae, M. roffensis, and M. martini.
• # 5. A single troglobiont in the northeastern Atlantic islands: Palmorchestia hypogea.

Of the five ecotopes in Table 4, the most variable one is the marine/estuarine supralittoral wrack zone. Wrack includes a wide range of different species of green, brown, and red macroalgae, driftwood (from natural tree fall and human discards), and may contain or be dominated by angiosperm leaf litter, including seagrasses such as Zostera sp. and Posidonia sp. Over the full range of a given wrack generalist species, for example, O. gammarellus from northern Africa to Norway, the composition of wrack is zoogeographically highly variable. In addition, each item of plant detritus in wrack may be at a different stage of the decay cycle. The amount of wrack at a given location varies temporally depending on tidal conditions (predictable) and wind effects (random and unpredictable). Thus, offshore winds may carry wrack and driftwood away as flotsam, while onshore winds may deposit new sources of wrack and driftwood. Because of the high degree of environmental variability, it is not surprising that species belonging to the wrack generalist ecotype are phenotypically plastic, e.g., in P. exter (as P. platensis, Stock 1996; LeCroy 2011), rendering morphological classification difficult.

Psammophilic talitrids (# 2 Table 4) are limited to sandy beaches into which they can make temporary burrows for residence during daylight hours. There are five species on the northwest Atlantic and six on the northeast Atlantic coasts. There is some evidence that psammophiles are selective feeders, choosing particular macroalgal species from wrack on which to feed. An example is Talitrus saltator in France, which prefers Fucus serratus over other wrack species (Adin and Riera 2003). At another location in Spain, T. saltator preferred two different species of brown (phaeophycete) seaweed wrack: Sargassum muticum and Gongolaria baccata (Olabarria et al. 2009). Such results suggest that availability (of a phaeophycete) within the local wrack determines the food choice. Most authors stress the importance of wrack aging (microbial decomposition) in food choice, thought to be because of increased nutritional value (Lastra et al. 2015) or the decomposition of grazer deterrent chemicals (phlorotannins) in brown macroalgae (Pennings et al. 2000).

Of the palustral ecotype (# 3 Table 4), there are possibly seven species on the northwest Atlantic coast, whereas in the northeast Atlantic, there are none. However, O. gammarellus may be present in salt marshes of the northeast Atlantic as a secondary palustral ecotype (see above, although experimental support that it is consuming seagrass leaf litter is lacking). Food available to salt marsh hoppers includes a range of salt marsh angiosperm plants, sometimes with the addition of marine macroalgae thrown into the marsh by storm winds or in flooding seawater. As far as I know, no studies on digestive physiology are available for talitrids where the primary ecotope is a salt marsh.

There are five known species of specialist xylophagous talitrids (# 4 Table 4) found only on the northeast Atlantic coast. They are obligate feeders on rotting driftwood. Recently, a new species of xylophagous talitrid was discovered in Thailand by Wongkamhaeng et al. (2022). It is the first driftwood hopper discovered outside the northeast Atlantic/Mediterranean region and is likely an example of convergent evolution.

The single species of cavernicolus talitrid (# 5 Table 4) occurs only in the northeastern Atlantic Islands, and its feeding habits have not been studied (as is true of all other troglobiontic talitrids so far discovered).

There are many islands present in the North Atlantic (International Hydrographic Organization 1953), of which only a few have been investigated to determine talitrid biodiversity. According to the influential model of island biogeography of MacArthur and Wilson (1967) and subsequent studies in insects (Peck and Finston 1993), the following factors might influence island occurrence by talitrids: island area, ecotope diversity on the island, the marine distance from the continental mainland, as well as island age. Factors affecting talitrid dispersal to oceanic islands include: emigration frequency from the donor population, the size of the donor population, and the dispersal capability of the transport raft (dependent on ocean current direction), including raft longevity (Wildish 2012).

As far as I am aware, only two oceanic islands have been studied sufficiently well to provide even a species list for the Talitridae occupying them. They are La Palma in the Canaries Archipelago in the eastern, and Bermuda in the western, North Atlantic.

The following types of macroevolution might occur during a study of the phylogeny of talitrids:

• Divergence: a once interbreeding population splits into two or more species, and the new species becomes more dissimilar with time. Many of the wrack generalist talitrids may have originated by divergent evolution.
• Convergence: occurs when two species from different genetic backgrounds occupy the same or similar ecotope and, by adaptative processes, come to show the same phenotypic characteristics. An example is the troglobiontic talitrids, in which all species so far recognized have lost eyes and epidermal pigment patterns.
• Parallel evolution: occurs when two species from genetically different ancestor lines develop similar phenotypic traits when occupying the same ecotope. An example would be the behavioral escape response of jumping in random directions to confuse avian predators shown by co-occurring Platorchestia oliveirae and Mexorchestia carpenteri carpenteri in the Gulf of Mexico wrack (Wildish and LeCroy 2014).
• Coevolution: when two often radically different species influence each other’s evolution to the benefit of both. An example is the wrack generalist Orchestia montagui feeding on wrack from Posidonia sp., which maintains a lignivorous bacterial species in the hepatopancreas capable of metabolizing the food that the talitrid eats (Abdelrhman et al. 2017). In return, the talitrid provides a suitable culture environment for the bacterium, which is rare or absent in the general environment.

A comprehensive fossil record for the Talitridae is absent. Currently, only three talitrid fossils have been discovered, two in Mexican amber (Bousfield and Poinar 1994; Hegna et al. 2020), the other from Dominican Republic amber (Bousfield and Poinar 1995), and all from 26–38 MYA. They were terrestrial species and are currently uninformative about their evolutionary history.

In classical taxonomy, a sufficient number of morphological characters must be chosen to distinguish between closely related populations and a subjective decision made by the taxonomist as to what constitutes a new species (Myers 1997). Such characters have also been used to infer the phylogeny of the Talitroidea. Thus, Bousfield (1982) used a phenetic analysis to do so within the Talitroidea, recognizing the problem of homoplasy and ecological overlap but apparently not correcting for them in his data matrix. More recently, Myers and Lowry (2020) have used a morphological cladistic analysis of the talitroids and determined for each character state whether the relationship was synapomorphic or homoplastic. A total of 46-character states were used in the analysis by Myers and Lowry (2020), of which 16 were zoogeographic in nature. Not surprisingly, there was little agreement between the two methods. A central problem with morphological reconstructions of amphipod phylogeny is that individual taxonomists use both a different number and quality of input data (morphological character states). Thus, in Bousefield’s (1982) phenetic analysis, 60 character states were included, involving some overlap with those used by Myers and Lowry (2020). The quality of the input data in the latter study was clearly higher than in the former.

No equivalent molecular-genetic method has been attempted to determine the phylogeny of the Talitroidea. The input data for such a study would be drawn from the genetic material found in each eukaryote cell. This method does offer one advantage over the morphological ones discussed above: it can be used to estimate the time that elapsed when a common ancestor branched into two clades using the molecular clock approach (Ho and Duchêne 2014). Smaller-scale phylogenies of Lake Baikal amphipods (Macdonald et al. 2005), world-wide species of Gammarus (Hou et al. 2007), and Ponto-Caspian amphipods (Mamos et al. 2021) have been based on molecular genetic data. A first attempt at a phylogeny of the Amphipoda (including Talitroidea) using a dataset containing representative species from approximately one-half of the known families of Amphipoda was published by Copilaş-Ciocianu et al. (2019). Four common genetic markers (mitochondrial CO1, nuclear ribosomal RNA subunits 28S and 18S, and nuclear histone H3) were used, and divergence time estimates for the Talitroidea were 40–72 MYA, depending on which of three calibration methods were used. This preliminary result is supportive of the more recent origin (70–110 MYA) of the Talitridae suggested by Bousfield (1984).

Ideally, we would test a data set common between a morphological cladistic analysis and one where molecular genetic data was available from the same taxa. Unfortunately, no such study has been attempted, and we must conclude that both methods are still under development and that such a test is premature. Thought experiments suggest that for the morphological cladistic analysis it must yield results that are similar when performed by different taxonomists who may be using different numbers and quality of morphological character states. Similarly, for molecular genetic input data, we would expect the same results if different molecular geneticists ran the analysis and chose different types of input data (mitochondrial or nuclear genes). For both approaches, independent scientists producing similar results would be a good indicator that they had achieved a reproducible scientific method.

In the North Atlantic, apart from the three amphi-Atlantic talitrid species of Table 2, all remaining 41 species are not shared between the northwestern and northeastern Atlantic coastlines. This is interpreted as circumstantial evidence of vicariance, in this case, the formation of the North Atlantic Ocean by the spreading apart of tectonic plates (Bousfield 1984; Myers 2022), which separated the North American and Eurasian continents. Thereafter, when the last land bridge between North America and Eurasia was severed at ~ 50 MYA (Atlantic Geoscience Society 2001), the fauna, including talitrids, evolved independently on the two separated continents.

One common route of land colonization by Talitridae has been demonstrated using knowledge of geological events in the Pacific Ocean (e.g., temporal patterns of sea level rise) and molecular genetic methods (Yang et al. 2013; Liu et al. 2022). The mechanism involved marine supralittoral talitrids in East Asia, which were carried inland by marine incursions when global sea level was high at 10–17 MYA.

Armed with the methods available to study ecology and microevolution, including field experimental ecological genetics (Ford 1975) coupled with modern molecular genetic methods and the new field of epigenetic evolution to explore, the student of these topics can make important discoveries with talitrids as the target. Both study areas are a good fit for young students of the Amphipoda keen to make fundamental discoveries with field biology methods. This is because so little has been achieved yet in either the ecology or evolution of the Talitridae. Another bonus for the aspiring talitrid student is that it will take you to interesting sampling locations.

We have seen in this review that the wrack generalist talitrids found in the Atlantic coastal region occupy an environment that offers the most variability: wrack, driftwood, salt marshes, and intertidal caves. Importantly, these are contemporary marine/estuarine ecotopes where each is adjacent or contiguous to each other and not solely dependent on passive dispersion to cross environmental boundaries. Passive mechanisms are thought to enable long-range dispersal (Wildish 2012). On the other hand, specialists occupy a single ecotope and have phenotypic adaptations specifically to it. The ecotopes of specialists become rarer as the continuum order becomes more specialized. Thus, caves that adjoin the marine intertidal are uncommon and far apart, rendering passive dispersal to and from them virtually impossible. These results emphasize the primacy of the environment in shaping the talitrid species that exist today.

Whether all wrack generalists share the ability to respond to the different ecotopes available remains to be demonstrated with those species not yet studied ecologically. A pertinent question is whether specialist ecotypes share the phenotypical plasticity characteristic given the experimental or natural opportunity to do so. The Palmorchestia example studied by Villacorta et al. (2008) suggests that the specialist does retain phenotypical plasticity to respond to contiguous biotopes (cave to subtropical forest leaf litter) in response to an undetermined environmental switch. Specialist ecotypes generally experience less environmental variability than generalists.

An important realization from this review is that the initial stages of talitrid microevolution may be controlled by epigenetic mechanisms in response to specific environmental factors and aided by switch genes. Two of the best examples given above may qualify as involving epigenesis. The first is the sibling pair O. mediterranea/O. aestuarensis, where the environmental cue is low salinity (Wildish and Radulovici 2020), and the second is the hypogaean/epigean sibling pair of Palmorchestia spp., with the environmental switch undetermined. Sibling pairs of hypogaean/epigean species are found in many other invertebrates, such as insects (Peck and Finston 1993), and the existence of sister species pairs could be a general indicator that epigenetic mechanisms were involved in their formation. If epigenesis is involved in microevolution in the formation of cryptic species complexes (Baldanzi et al. 2016, 2017), it has implications for morphological systematics as currently practiced in the Talitridae (Myers and Lowry 2020, 2023). As we have seen, wrack generalist talitrids are biologically vagile, including in their morphology in responding to the ecotope variability that they meet throughout their zoogeographic range. Such a situation poses a problem for the biological species concept (Mayr 1942) and the morphological systematics of talitrids. Part of the solution involves the introduction of molecular genetic methods into the routine practice of systematics (Coleman and Radulovici 2020), as well as increased field sampling to discover new talitrids.

For variation in dorsal pigment patterns used as camouflage protection against avian predators, the genetic variation brought about by sexual reproduction, including crossing over during meiosis (Jablonka and Lamb 2014), may be key. Here, natural selection by bird predation at the population level can determine which pigment patterns survive against the background ecotope in which a population lives. For talitrid dwarfness, two environmental drivers have been identified. The first involves wrack generalists, such as Platorchestia exter, who, if they adopt driftwood as food, have a reduced standard metabolic rate, slower growth, and final body size, and if the usual wrack diet is offered, metabolism and growth rate are restored (Wildish and Robinson 2016a). The second case involves obligate driftwood-living talitrids of the genus Macarorchestia. The five species of this genus form a series of decreasing sizes, with the leading hypothesis being that they were adapting to gribble burrow hole diameters in driftwood (Wildish and Robinson 2016b).

Is it better to become a specialist or remain a generalist, that is, to specialize or not to specialize? The question is rhetorical and does not imply that the talitrid has a choice. The business end of the matter is the individual and population undergoing natural selection in response to the environmental limiting factors it has to face. The latter include those environmental limits involved in passive dispersal mechanisms of talitrids. Comparing the wrack generalist with the troglobiont (Table 4), the main difference between them is that there are many more individual locations available for the former than the latter. This results from the larger zoogeographic range of wrack generalists and therefore more nuanced ecotope variations to which it is exposed. The collateral effects that follow (Table 8) are that phenotypic variation is higher, the zoogeographic range is larger, speciation potential is lower, and endemism is lower in wrack generalists than in talitrid specialists. This implies the primacy of environmental variability in determining whether to speciate or not to speciate.

The following are thanked for reviewing earlier versions: John McDonald, Gerhard Pohle, and Davide Asnicar. Alan Myers and Luiz de Andrade improved a later version.