Desiccation tolerance of the introduced marine green alga <Emphasis Type="Italic">Codium fragile</Emphasis> ssp. <Emphasis Type="Italic">tomentosoides</Emphasis> – clues for likely transport vectors?

The invasive marine green macroalga Codium fragile ssp. tomentosoides is now considered to be an introduced marine pest along the northwest Atlantic and southern Australian coasts. International or domestic translocation of C. fragile ssp. tomentosoides is usually attributed to the fouling of ship hulls or shellfish, particularly oysters. A likely domestic vector is shipboard transport, involving the translocation of whole thalli or fragments entangled in fishing nets, ropes, etc. that are then released in a previously unaffected area. Here we investigated the survival of C. fragile ssp. tomentosoides under emersed conditions, simulating shipboard transport. C. fragile ssp. tomentosoides was able to survive periods of emersion of up to 90 days in high relative air humidity. Net photosynthesis remained positive at about 50% of the rates of submersed control thalli. After 2 days of emersion and 4 days of rehydration under submersed conditions thalli recover to their initial net photosynthesis rates. Hence, C. fragile ssp. tomentosoides is likely to survive long shipboard journeys entrapped in fishing nets, anchor wells or other protected, high-humidity areas of a vessel. Furthermore, C. fragile ssp. tomentosoides may survive emersion on an exposed deck during short trips, especially in cooler conditions such as at night. The incursion sites of C. fragile ssp. tomentosoides in Australia are generally in modified environments, often associated with shipping-related infrastructure such as wharves, jetties, rip rap, and moorings.     

Temperature,Herbivory and Epibiont Acquisition as Factors Controlling the Distribution and Ecological Role of an Invasive Seaweed

The invasive canopy alga, Codium fragile ssp. tomentosoides, first observed at the Isles of Shoals in 1983, has become the dominant canopy species to 8 m throughout the islands. Codium populations are replacing themselves at most sites in what appears to be a new, climax, canopy species. However, Codium densities have declined in protected Gosport Harbor areas where it first became established. Codium has only slowly expanded its presence in adjacent nearshore subtidal habitats. Recent studies suggest a combination of factors that may be influencing the relative success of populations between habitats. The herbivorous sea slug, Placida dendritica, may be reducing populations in protected areas in spite of predators such as the green crab, Carcinus maenas, while surge may inhibit herbivore buildup in exposed habitats. Temperature instability due to localized, wind-driven upwelling may be slowing the buildup of subtidal Codium populations in nearshore sites. The combination of Codium dominance and the acquisition of increasing epibiont diversity are producing a new, potentially more complex community state than the previous kelp-dominated climax typical of the Gulf of Maine.     

Dispersal potential of the invasive green alga Codium fragile ssp. fragile

Since its introduction in 1989, the invasive green alga Codium fragile ssp. fragile (formerly Codium fragile ssp. tomentosoides) has spread rapidly in Atlantic Canada. Although its spread has likely been facilitated by human transport, C. fragile possesses diverse modes of natural dispersal. In addition to parthenogenetically developing swarmer cells, it can propagate through the release of vegetative buds, thallus fragments, and entire, dislodged thalli. We examined the natural dispersal potential of these propagules using a combination of field and laboratory experiments. Vegetative buds were most abundant on thalli in late August, coinciding with increasing late summer wave activity. At two locations within two field sites, we examined the effects of tidal state, wave action and topography on the dispersal of fragments and intact thalli. Over a 4-hour period, retention rate was generally lower and dispersal distance of retained propagules higher: during flood than ebb tide; at the more wave-exposed locations; and for fragments than thalli. Dispersal direction corresponded with topography or flood tide currents at certain locations, but was often bi-directional or random, suggesting an added role of wave action. Over periods of weeks, the retention rate of marked detrital thalli was negatively related to the magnitude of wave activity. In the laboratory, settling rate was negatively related to propagule length, and large buds had higher critical shear erosional velocities than either small buds or fragments. Our results indicate that large propagules of C. fragile, such as intact adult thalli, have a generally low dispersal potential (meters), but may be transported longer distances (kilometers) during storms or when positively buoyant. Fragments and buds are regularly transported 10's of meters at a time at average flow velocities. However, they also become resuspended in average flows, and probably disperse longer distances over multiple resuspension events, or when positively buoyant. Once settled, the smallest propagules may be less easily resuspended or transported along the bottom as they become entrapped by small-scale topographic features or turf algae. The wide variety of propagules produced by C. fragile and the variable distances over which different propagule types may be transported give C. fragile the advantage of both short- and long-distance dispersal, and have likely played a role in the invasive success of this alga.     

<Emphasis Type="Italic">Codium fragile</Emphasis>: rhizomatous growth in the<Emphasis Type="Italic"> Zostera</Emphasis> thief of eastern Canada

A rhizomatous growth form of Codium fragile is described for the first time. Plants were collected in the Gulf of St. Lawrence in estuaries dominated by Zostera marina. Rhizomatous plants developed from propagules of whole plants that settled horizontally. Horizontal axes of C. fragile were up to 1 m long in plants collected in situ. Plants developed several to dozens of erect axes at right angle to the base. Horizontal growth of up to 0.2 m was found in field experiments where fragments were tied to plastic mesh and left in situ for 4 months. The unconsolidated filaments at the base of C. fragile often wrapped around the rhizomes of Z. marina and up to five separate attachment sites to eelgrass were found in single plants of C. fragile. In four estuaries, 57–100% of Codium plants with identifiable substratum were attached to shoots and rhizomes of Z. marina. The rhizomatous growth form was found in plants identified as C. fragile ssp. tomentosoides (Nova Scotia and Prince Edward Island) and C. fragile ssp. atlanticum (Prince Edward Island), suggesting that this is a phenotypic response to growth in soft bottom environments.Communicated by K Lüning     

TRACING SPECIES INVASION IN CODIUM,A SIPHONOUS GREEN ALGA,USING MOLECULAR TOOLS

The siphonous green alga Codium fragile occurs in many temperate marine regions and is composed of a number of distinct subspecies. Included in this taxon is the common open coast C. fragile subsp. fragile of the northeast Pacific and the weedy C. fragile subsp. tomentosoides which has invaded temperate marine communities in the Atlantic and Pacific oceans. The center of origin of this weedy subspecies is not known, although it is thought to have dispersed from the northwest Pacific. To examine the relationship of the weedy subspecies to the indigenous northeast Pacific form, chloroplast DNA was compared. Each of these subspecies has a restriction map that is uniform throughout its geographic distribution, and the patterns are distinct from each other and from other Codium species examined. However, the two share an almost identical genome size and arrangement of genes. A population in San Francisco Bay was found to be indistinguishable from the weed C. fragile subsp. tomentosoides from the Atlantic. The potential for using molecular data in solving systematic problems in Codium has been clearly demonstrated.     

Grazing of the invasive alga Codium fragile ssp. tomentosoides by the common periwinkle Littorina littorea: Effects of thallus size, age and condition

We examined the potential of herbivory by the common periwinkle Littorina littorea to limit recruitment and vegetative re-growth of the invasive green alga Codium fragile ssp. tomentosoides in a series of manipulative field experiments in tidepools on a wave-exposed rocky shore in Nova Scotia, Canada. Snails were excluded or included from circular plots (14 to 20 cm diameter) with cages to compare growth and survival of C. fragile against procedural (partial cages) or natural (uncaged) control plots. Our results show that L. littorea may restrict growth and survival of C. fragile by grazing new recruits (< 2 cm thallus length), fronds of adult thalli that are bleached and necrotic, and residual holdfasts (< 2 mm thickness) of detached thalli (artificially severed to mimic wave dislodgement). Once recruits grow beyond a critical size (~ 3 cm), or damaged tissues or holdfasts regenerate, grazing of C. fragile by L. littorea appears to be limited and ineffectual. Our experimental results corroborate correlative evidence from previous studies that herbivory by L. littorea limits the abundance of C. fragile in tidepools on the Atlantic coast of Nova Scotia, particularly pools in the high intertidal zone where these snails are abundant. Lower on the shore, littorinid grazing and physical stressors that render algae more vulnerable to grazers (e.g. UV radiation and freezing) are less intensive, and probably have less of a regulatory effect on populations of C. fragile.     

Changing Community States in the Gulf of Maine: Synergism Between Invaders, Overfishing and Climate Change

Human activities, including overfishing and species introductions, have had a dramatic impact on benthic communities in the Gulf of Maine within the past two decades. Prior to the 1970s, the climax community in the shallow subtidal was composed of Laminaria spp. kelp beds with an understory of arborescent red algae. In the 1980s, a population explosion of the green sea urchin, Strongylocentrotus droebachiensis, created an alternate community state, urchin barrens. Recently, a new community has been observed in former urchin barrens and kelp beds. This assemblage is principally composed of the introduced species: Codium fragile subsp. tomentosoides (green alga), Membranipora membranacea (bryozoan), Diplosoma listerianum (tunicate), Bonnemaisonia hamifera (red alga) and the opportunistic species Mytilus edulis (mussel) and Desmarestia aculeata (brown alga). In addition to changes in relative abundance, many of these species have greatly expanded their distribution and habitat selection. A model detailing mechanisms for the transition of the traditional kelp bed and urchin barren communities to others is presented and implications for this new community are discussed. This revised version was published online in July 2006 with corrections to the Cover Date.     

Occurrence of a fouling,juvenile, stage ofCodium fragile ssp.tomentosoides (Goor) Silva (Chlorophyceae,Codiales)

An unusual, siphonaceous, green algal thallus grows on the support floats of marinas in the Solent area of southern England. Comparative studies of British species of the order Codiales collected from the marinas, particularly relating to vegetative structure, ecology, betaine and tertiary sulphonium compounds and lectin activity of extracts of the algae, led to the conclusion that the thallus was a juvenile stage ofCodium fragile ssp.tomentosoides.     

Spray system for re-attachment of Gelidium sesquipedale (Clem.) Born. et Thur. (Gelidiales: Rhodophyta)

A system is described for rapid re-attachment of the rhodophycean alga Gelidium sesquipedale (Clem.) Born. et Thur. on artificial or natural substrata. This method is applicable to industrial cultivation of this species. The function of rhizoidal clusters and the origin of germlings from the apical portion of the thalli are analyzed in relation to the re-attachment process. The role that re-attachment might play in the maintenance and spreading of natural populations and in the observed anomalies of the life-history of this species is discussed.     

Cultivation of the green alga, Codium fragile (Suringar) Hariot, by artificial seed production in Korea

Codium fragile (Suringar) Hariot is an edible green alga farmed in Korea using seed stock produced from regeneration of isolated utricles and medullary filaments. Experiments were conducted to reveal the optimal conditions for nursery culture and out-growing of C. fragile. Sampling and measurement of underwater irradiance were carried out at farms cultivating C. fragile at Wando, on the southwestern coast of Korea, from October 2004 to August 2005. Growth of erect thalli and underwater irradiance were measured over a range of depths for three culture stages. During the nursery cultivation stage (Stage I), growth rate was greatest at 0.5 m depth (0.055 ± 0.032 mm day⁻¹), where the average midday irradiance over 60 days was 924 ± 32 μmol photons m⁻² s⁻¹. During the pre-main cultivation stage (Stage II), the greatest growth rate occurred at a depth of 2 m (0.113 ± 0.003 mm day⁻¹) with an average irradiance of 248 ± 116 μmol photons m⁻² s⁻¹. For the main cultivation stage (Stage III) of the alga, thalli achieved the greatest increase in biomass at 1 m depth (7.2 ± 1.0 kg fresh wt m⁻¹). These results suggest that optimal growth at each cultivation stages of C. fragile could be controlled by depth of cultivation rope.     

Mismatch of morphological and molecular identifications in native and invasive subspecies of Codium fragile (Bryopsidophyceae,Chlorophyta)

Several subspecies are defined within Codium fragile, including the invasive C. fragile ssp. fragile, first reported in New Zealand in 1973. An endemic subspecies, C. fragile ssp. novae‐zelandiae, is also found throughout New Zealand. The two subspecies exhibit morphological and molecular variation, although these have never been evaluated together. We compared variation between subspecies at locations in Auckland, identifying subspecies using rps3‐rpl16 DNA sequence data, and assessing gross morphological differences, anatomical utricle characters and morphometrics. The taxonomic utility of the morphometric data sets was assessed by linear discriminant analysis. Utricle characters and measurements varied within individual thalli and between different preservation methods. The phenotypes of both subspecies were highly variable and influenced by environment. Accurate subspecies delimitation using morphological data was not possible; the discriminant analyses performed no better than chance for all combinations of the morphological data. Specimens from New Zealand, Canada, Australia and Ireland were sequenced using both the rps3‐rpl16 and tufA plastid markers. The tufA elongation factor was shown to be a good candidate for differentiating subspecies of C. fragile. This marker is twice the length of the rps3‐rpl16 spacer, shows greater variation between ssp. fragile and novae‐zelandiae, and is less prone to sequencing error. A simple restriction enzyme digest of the tufA amplicon can distinguish ssp. fragile and ssp. novae‐zelandiae. Our study expands the known range of the ssp. fragile in New Zealand, including the first record of this subspecies from the west coast of Auckland, and points to a need to re‐evaluate morphological and molecular criteria for subspecies currently defined within C. fragile.     

Invasion of alien macroalgae in different Mediterranean habitats

The study evaluated different macroalgal invasions in the main Mediterranean coastal habitats on hard bottom. Biodiversity, species composition and structure of macroalgal assemblages were compared among non-invaded areas and areas invaded by the Chlorophyta Caulerpa racemosa var. cylindracea and by the turf-forming Rhodophyta Womersleyella setacea in three different habitats: shallow rocky bottom, deep rocky bottom and dead matte of the seagrass Posidonia oceanica. Results showed that alien macroalgae constituted a relevant component of benthic assemblages in invaded areas of the Mediterranean Sea. Assemblages invaded by Womersleyella setacea and Caulerpa racemosa showed lower values of diversity and large differences in the structure and species composition related to non-nvaded assemblages. The species that mostly suffered from invasion were erect species reproducing sexually; moreover, the dominance of W. setacea led to low abundance of native filamentous algae, while C. racemosa colonization seemed particularly threatening for encrusting algae. All the studied habitats appeared highly invasible by alien macroalgae, even if W. setacea appeared more invasive in deeper habitats, while colonization of C. racemosa seemed more serious in shallower habitats; the dead matte of P. oceanica represented a suitable substrate for the spread of both species. Differences among assemblages in different habitats were reduced in invaded areas.     

Contrasting patterns of spread in interacting invasive species: Membranipora membranacea and Codium fragile off Nova Scotia

In the Northwest Atlantic, overgrowth of the competitively dominant, native kelps by an invasive bryozoan Membranipora membranacea increases frond erosion, which has facilitated the establishment and spread of the invasive macroalga Codium fragile ssp fragile. To document the spread of both introduced species along the Atlantic coast of Nova Scotia from initial introduction points (the ‘epicentre’) southwest of Halifax, we conducted video-surveys of shallow rocky habitats along the southwestern shore of Nova Scotia (100 km linear distance, encompassing the range of M. membranacea) in 2000, and then along the entire Atlantic coast in 2007 (650 km). Membranipora membranacea was observed continuously throughout the surveyed ranges in 2000 and 2007, wherever kelps were present, suggesting natural dispersal via planktonic larvae. Codium fragile was observed along 95 km of the surveyed range in 2000 and along 445 km in 2007, with a relatively patchy distribution beyond the epicentre, suggesting a combination of natural and anthropogenic dispersal mechanisms. Rockweed-dominated (Fucus spp.) or mixed algal assemblages common outside the epicentre may alter the interaction between M. membranacea and C. fragile, since seaweeds other than kelp are not subject to defoliation by the bryozoan. Percent cover of kelp at the epicentre generally increased from 2000 to 2007, while that of C. fragile generally decreased. Codium fragile was the dominant canopy alga at 54% of sites in 2000 and at only 15% of sites in 2007. These findings indicate that, at near decadal timescales, C. fragile does not prevent re-colonization by native kelps.     

Aggregation dynamics and foraging behaviour of striped red mullet Mullus surmuletus in the western Mediterranean Sea

Habitat‐related heterogeneity of striped red mullet Mullus surmuletus heterospecific foraging assemblages was examined off the coast of Spain. Video‐based focal‐follows conducted on 122 M. surmuletus assemblages (446 total individuals) revealed an array of attendant species (n = 7) with composition linked to benthic habitat complexity; bare sandy substrata were characterized by homospecific groups of M. surmuletus, while habitats with rock and vegetation attracted a variety of scrounging labrids and sparids. Although the nature of the relationship between M. surmuletus and attendants requires further exploration, the present study indicates that substratum composition can be a driving factor explaining the dynamics of this heterospecific assemblage.     

Introduced Macroalgae – a Growing Concern

Introductions of non-indigenous species to new ecosystems are one of the major threats to biodiversity, ecosystem functions and services. Globally, species introductions may lead to biotic homogenisation, in synergy with other anthropogenic disturbances such as climate change and coastal pollution. Successful marine introductions depend on (1) presence of a transport vector, uptake of propagules and journey survival of the species; (2) suitable environmental conditions in the receiving habitat; and (3) biological traits of the invader to facilitate establishment. Knowledge has improved of the distribution, biology and ecology of high profile seaweed invaders, e.g. Caulerpa taxifolia, Codium fragile ssp. tomentosoides, Sargassum muticum, and Undaria pinnatifida. Limited, regional information is available for less conspicuous species. The mechanisms of seaweed introductions are little understood as research on introduced seaweeds has been mostly reactive, following discoveries of introductions. Sources of introductions mostly cannot be determined with certainty apart from those directly associated with aquaculture activities and few studies have addressed the sometimes serious ecological and economic impacts of seaweed introductions. Future research needs to elucidate the invasion process, interactions between invaders, and impacts of introductions to support prevention and management of seaweed introductions.     

Detecting the impacts of notorious invaders: experiments versus observations in the invasion of eelgrass meadows by the green seaweed <Emphasis Type="Italic">Codium fragile</Emphasis>

Biological invasions can vary in the extent of their effects on indigenous communities but predicting impacts for particular systems remains difficult. In coastal marine ecosystems, the green seaweed Codium fragile ssp. fragile is a notorious invader with its reputation based on studies conducted largely on rocky shores. The green seaweed has recently invaded soft-bottom eelgrass communities by attaching epiphytically to eelgrass (Zostera marina) rhizomes, thereby creating the potential for disruption of these coastal habitats through competition or disturbance. We investigated the effect of this invader on various aspects of eelgrass performance (shoot density and length, shoot growth, above- and below-ground biomass, carbohydrate storage) using both small-scale manipulative and large-scale observational experiments. Manipulative experiments that varied Codium abundance demonstrated clear negative effects over a 4-month period on shoot density and carbohydrate reserves, but only for high, but realistic, Codium biomass levels. Light levels were much lower under canopies for high and medium density Codium treatments relative to low and control Codium cover treatments, suggesting that shading may influence eelgrass growing under the algal cover. In contrast, these effects were either not detectable or very weak when examined correlatively with field surveys conducted at larger spatial scales, even for sites that had been invaded for over 4 years. It is premature to extend generalizations of Codium’s impact derived from studies in other systems to eelgrass communities; further efforts are required to assess the long-term threats that the alga poses to this ecosystem. This study demonstrates the need to investigate impacts of invasions over multiple scales, especially those that incorporate the temporal and spatial heterogeneity of the invader’s abundance.     

From introduced species to invader: what determines variation in the success ofCodium fragile ssp.tomentosoides (Chlorophyta) in the North Atlantic Ocean?

The green algaCodium fragile ssp.tomentosoides (Chlorophyta) has been introduced accidentally and successfully from Japan to many shores of the northern and southern hemispheres, including those of the Northeast and Northwest Atlantic Ocean. On most European coasts,Codium occurs regularly but at low abundances in the intertidal zone and is absent from subtidal habitats. In contrast,Codium is extremely abundant in subtidal kelp beds in the Northwest Atlantic Ocean where it often reaches nuisance proportions. This differential success cannot be accounted for by either the properties of the invader or by physico-chemical differences between invaded coasts. A theoretical comparison between two regions on opposite sides of the Atlantic Ocean, i.e. Eastern Nova Scotia, Canada, and south central Britain, illustrates how the resident benthic community may determine the difference in relative abundance ofCodium in subtidal habitats between northeast America and Europe. In this review, low floral species diversity, biological disturbance and facilitation by a previous species invasion are suggested as potential factors for the establishment, success and abundance ofCodium in the Northwest Atlantic Ocean, but these require testing in field experiments.     

Anthropogenic disturbance can determine the magnitude of opportunistic species responses on marine urban infrastructures

Background

Coastal landscapes are being transformed as a consequence of the increasing demand for infrastructures to sustain residential, commercial and tourist activities. Thus, intertidal and shallow marine habitats are largely being replaced by a variety of artificial substrata (e.g. breakwaters, seawalls, jetties). Understanding the ecological functioning of these artificial habitats is key to planning their design and management, in order to minimise their impacts and to improve their potential to contribute to marine biodiversity and ecosystem functioning. Nonetheless, little effort has been made to assess the role of human disturbances in shaping the structure of assemblages on marine artificial infrastructures. We tested the hypothesis that some negative impacts associated with the expansion of opportunistic and invasive species on urban infrastructures can be related to the severe human disturbances that are typical of these environments, such as those from maintenance and renovation works.

Methodology/Principal Findings

Maintenance caused a marked decrease in the cover of dominant space occupiers, such as mussels and oysters, and a significant enhancement of opportunistic and invasive forms, such as biofilm and macroalgae. These effects were particularly pronounced on sheltered substrata compared to exposed substrata. Experimental application of the disturbance in winter reduced the magnitude of the impacts compared to application in spring or summer. We use these results to identify possible management strategies to inform the improvement of the ecological value of artificial marine infrastructures.

Conclusions/Significance

We demonstrate that some of the impacts of globally expanding marine urban infrastructures, such as those related to the spread of opportunistic, and invasive species could be mitigated through ecologically-driven planning and management of long-term maintenance of these structures. Impact mitigation is a possible outcome of policies that consider the ecological features of built infrastructures and the fundamental value of controlling biodiversity in marine urban systems.     

Effects of <Emphasis Type="Italic">Caulerpa racemosa</Emphasis> invasion on soft-bottom assemblages in the Western Mediterranean Sea

The present study aimed at evaluating the effects of Caulerpa racemosa var. cylindracea spread on Mediterranean soft-bottom assemblages. The benthic assemblages colonized by C. racemosa were compared with non invaded assemblages at multiple spatial scales. In addition, a manipulative experiment has been conducted over a one-year period in order to compare the structure of native assemblages invaded by the alga with others where the alga has been manually removed and others that were not invaded. Results of both studies showed that Mediterranean soft-bottom assemblages invaded by C. racemosa differed from non invaded ones in terms of species composition, abundance and patterns of spatial variability. Moreover, in areas cleaned out from the alga, the benthic assemblages begin to recover their structure and after one year they get more similar to the assemblages observed in non invaded areas. The present paper, taking into account different aspects of C. racemosa invasion in Mediterranean soft bottoms, highlighted that the observed increase in alpha diversity did not correspond to an increase in the overall diversity of the studied system. Infact, the loss of beta diversity in invaded sites, together with the spread of species typical of vegetated habitats, may contribute to a homogenization of the Mediterranean coastal system. Furthermore, a possible role of ecosystem engineer for C. racemosa can be hypothesized.     

Re-establishment of epibiotic communities in reforested mangroves of Gazi Bay, Kenya

Recolonization of epibiotic flora and fauna in two fringing Sonneratia alba reforestation plots was investigated and compared to a natural mangrove stand and a denuded site in Gazi Bay, Kenya. The reforested sites differed with respect to land history and planting density. Habitat availability in the form of pneumatophore surface differed among forested sites (P<0.001), and between landward and seaward zones (P<0.05). Eighteen algal species were found in the natural area compared to 23 and 10 in replanted sites. Only one species was encountered in the denuded area. SIMPER analysis distinguished Enteromorpha ramulosa, Polysiphonia sp., Hypnea sp. and Caloglossa leprieuri as the main algal species responsible for differences between sites. Algal biomass was positively correlated to pneumatophores area (P<0.001). Total algal biomass differed markedly between forested sites: 1.4 (matrix replantation), 28.6 (natural stand) and 44.3 g m⁻² (integrated replantation) in the seaward zones. The matrix replantation showed strong differences in algal community assemblages compared to the other forested sites, and this site also had significantly lower biomass of sessile benthic fauna (P<0.001). Statistical differences in algal (P<0.01) and sponge (P<0.05) community composition between landward and seaward zones were observed in all sites and trunk fouling fauna was distinctly different between sites. Reasons for the above patterns are discussed and it is suggested that zonation patterns affecting pneumatophore surface and inundation time, in combination with proximity of sites to natural seeding areas, are the most likely explanations for observed patterns of epibiotic community distribution in this study.