There is currently no broad consensus regarding the systematic classification of Eurasian and North African brown trouts, an assemblage comprising all representatives of the genus Salmo except the well-differentiated Atlantic Salmon (Salmo salar), Marble Trout (Salmo marmoratus), Softmouth Trout (Salmo obtusirostris) and Ohrid Belvica (Salmo ohridanus). While numerous, often range-restricted, members of this grouping have been described based largely on their ecological and morphological diversity, this variability is not consistently reflected by phylogenetic and phylogeographic evidence (Sanz 2018; Whiteley et al. 2019; Segherloo et al. 2021).

Despite a relatively recent diversification history spanning the period 0.5-2.5 Mya, brown trouts exhibit marked ecological and phenotypic variability throughout their large native range, which extends eastward from Europe and Northwest Africa to Russia and the Aral Sea basin. They occupy a wide range of habitats, from mountain streams and larger rivers to lakes and estuaries. Individual subpopulations can exhibit sedentary, anadromous or potamodromous life history strategies. Some freshwater systems are inhabited by multiple sympatric forms which differ in traits associated with foraging and reproductive ecology, and are sometimes referred to as "morphs", "ecomorphs" or "ecotypes" (Klemetsen et al. 2003, Kottelat and Freyhof 2007, Ferguson et al. 2019, Segherloo et al. 2021).

Some authorities have viewed this combination of factors to be representative of high species diversity and recognised around 50 nominal taxa, a number of which have been described this century (Kottelat and Freyhof 2007, Snoj et al. 2011, Sanz 2018). Alternatively, their systematics have been viewed from a phylogenetic and phylogeographic perspective based largely on mitochondrial DNA (mtDNA) analyses, with all subpopulations treated as a single polymorphic taxon customarily referred to as the “Brown Trout (Salmo trutta) complex” (Sanz 2018, Whiteley et al. 2019, Segherloo et al. 2021).

The latter approach led to brown trout diversity being defined by ten mtDNA lineages or sublineages corresponding to extensive catchments (the Danube, Atlantic, Mediterranean and Adriatic basins), specific geographic areas (the Balkan Peninsula and North Africa), individual watersheds (the Dades, Duero and Tigris rivers) and a distinctive phenotype (Marble Trout). Subsequent studies revealed that the distribution of some of these mtDNA lineages extends beyond their defined boundaries, e.g., the Adriatic lineage occurs from the Iberian Peninsula to the Republic of Türkiye, and the Marble Trout lineage is present in areas where no marbled phenotype exists, such as Corsica, central Italy, Albania and Greece (Bernatchez et al. 1992, Apostolidis et al. 1997, Bernatchez 2001; Suárez et al. 2001, Cortey and García-Marín 2004, Sušnik et al. 2005, 2007; Splendiani et al. 2006, Martínez et al. 2007, Snoj et al. 2009, 2011; Tougard et al. 2018, Schöffmann et al. 2022).

However, several studies have revealed the presence of mosaic distributions of mtDNA haplogroups among wild brown trout populations, plus mitochondrial-nuclear phylogenetic discordance in reconstructions made with both mitochondrial and nuclear trees (Snoj et al. 2009, Pustovhr et al. 2014, Leucadey et al. 2018; Splendiani et al. 2020). This suggests the presence of incomplete lineage sorting or asymmetric introgressive hybridization, which are common phenomena in rapidly diverging lineages and indicate that mtDNA genealogies might be generally unsuitable for defining phylogenetic relationships between brown trout taxa (Pustovhr et al. 2011, 2014). In the case of brown trouts, naturally intricate patterns of diversification and secondary contact shaped by repeated glaciations during the Pleistocene have been additionally complicated by widespread anthropogenic translocation and introgressive hybridisation since the Middle Ages (Largiadèr and Scholl 1996, Sanz et al. 2006, Lerceteau-Köhle et al. 2013). The combined use of multiple nuclear (nDNA, e.g., microsatellites, nuclear genes) and mitochondrial markers has already provided better insight into this complex scenario, resulting in progress towards a deeper understanding of evolutionary relationships at particular geographic scales or among subsets of putative taxa (Snoj et al. 2002, 2010, 2011; Sušnik et al. 2006, 2007; Berrebi et al. 2013, 2019; Gratton et al. 2014, Marić et al. 2017).

An integrative taxonomic approach combining morphological and ecological data with next generation sequencing of nDNA to identify genomic clusters may represent the most promising option for resolving brown trout systematics (Guinand et al. 2021; Segherloo et al. 2021). However, no comprehensive morphological or nDNA analyses have yet been completed, and it is plausible that the elaborate genetic and phenotypic diversity demonstrated by these fishes may never be adequately captured by a single accepted taxonomic system (Whiteley et al. 2019).

Pending a definitive outcome to the above, the Red List broadly follows the nomenclature provided by Fricke et al. (2024).

The Garda Trout has a restricted range (extent of occurrence (EOO) c. 654 km², area of occupancy (AOO) c. 368 km²), which meets the thresholds for the Endangered category under Criterion B1 (EOO < 5,000 km²) and Criterion B2 (AOO < 500 km²). It is present at one location where the quality of habitat is estimated to be declining. Therefore, this species is assessed as Endangered under Criterion B (B1ab(iii)+2ab(iii)).

This species is endemic to Lake Garda in the Mincio River system, northern Italy. The Mincio is a major tributary of the Po River.

It has in the past been introduced to a number of other lakes in Italy and elsewhere, e.g., Germany, New Zealand, but did not become established in any of them.

This species' current population size and trend have not been quantified, but it has suffered a dramatic population size reduction since the mid-20^th century.

The pattern of decline is demonstrated by fisheries data, which show that commercial landings collapsed from an average of 20 tonnes per year during the 1950s to fewer than 5 tonnes by the 1970s. Between the 1990s and mid-2010s the average yield was < 2 tonnes.

In terms of genetic structure, this species is included in the Adriatic mitochondrial lineage within the Brown Trout (Salmo trutta) complex, but is understood to be naturally introgressed with the Marble lineage (see 'Taxonomic Notes').

Garda is Italy's largest natural lake and is a deep (maximum depth 346 metres, with the deepest point c. 280 m below sea level), oligomictic, naturally oligotrophic system formed after the last glacial period. It has a comparatively long retention time due to its volume and relatively low output through the Mincio River, and full mixing of the water column occurs only during periodically harsh winters. The lake experienced anthropogenic nutrient enrichment during the 20^th century, and is currently classified as oligo-mesotrophic (see 'Threats').

The Garda Trout occupies bentho-pelagic habitats in the profundal zone at depths of 100-200 metres, where it feeds on benthic invertebrates, zooplankton and occasionally smaller fishes. Unlike related species, it is markedly gregarious and is understood to exhibit schooling behaviour throughout the year.

The maximum recorded age is 5+, and individuals mature at age 2-4+. There are two annual reproductive periods of which the longer extends from December to February, with a second brief peak in July and August. Molecular analyses have demonstrated that this phenomenon is not indicative of population structuring or reproductive isolation, and at least some individuals participate in spawning during both of these periods. Spawning takes place within the lake itself, with the gametes deposited on beds of well-washed coarse stony substrata at depths of 50-200 metres. Known winter spawning sites are located along Lake Garda's northwestern shore and summer sites in the southern part of the lake, mainly in the vicinity of river mouths, debris from landslides or submerged ridges. Nuptial male individuals develop a conspicuous epigamic colour pattern comprising darkened fins and a more intense base body tone.

Molecular analyses have demonstrated that this species has not suffered from introgressive hybridisation with non-native Brown Trout (Salmo trutta), which was probably introduced to the Lake Garda system around the turn of the 20^th century. This is probably the result of reproductive isolation due to its selection of deep-water spawning sites that are inaccessible to non-native congeners.

The precise cause of this species' decline remains somewhat unclear, but it has plausibly been driven by multiple factors.

At least nine non-native fish taxa are established in Lake Garda, among which resource competition with planktivorous whitefish (Coregonus spp.) has often been cited as a primary threat to the Garda Trout. Whitefish stocking fluctuated somewhat following its introduction in 1918, but generally increased after the 1950s and has stabilised at an average of 40 million fry per year since the mid-2000s. Whitefish are currently the most important component of the lake's commercial fishery, with an annual yield of 100-250 tonnes per year.

In addition, the benthic Burbot (Lota lota) has been present in the lake since the 1850s and may predate on Garda Trout eggs and larvae. The invasive Pontocaspian Zebra Mussel (Dreissena polymorpha) was first reported during the 1960s, and the related Quagga Mussel (Dreissena bugensis) in 2022. The impact of these filter-feeding molluscs on the lake's ecosystem have not been investigated, but in some Swiss perialpine lakes they are understood to be driving significant food web alterations, including energy sources and pathways for native fishes, due to their propensity to alter zooplankton abundance, community structure and composition.

The introduction of industrial harvesting methods during the mid-20^th century (see 'Use and Trade') might also have contributed to declining Garda Trout abundance. Prior to the 1980s, commercial fishing also took place during the two annual reproductive periods and a combination of these factors may have driven a progressive reduction in the number of spawning individuals.

Since 1920, Lake Garda has been extensively modified and converted into a multipurpose reservoir which includes a number of hydropower and irrigation schemes. Among these, completion of the Salionze Dam on the upper Mincio River in 1951 altered the relationship between the lake's discharge and its surface level, and hydroelectric projects on the affluent Sarca River have interfered with the seasonal inflow regime. At the northeastern end of the basin, the Adige-Garda diversion tunnel is an inter-basin transfer system that directs water from the adjacent Adige River system to the lake during flood events, and may have caused a reduction in the extent and quality of Garda Trout spawning sites through increased sedimentation. The littoral zone of the lake has been extensively degraded by construction of urban and recreational infrastructure, which has led to the loss of macrophytes and reedbeds and is likely to have further increased erosion alongside increased pressure from tourist boats.

During the 20^th century, rising agricultural, industrial, urban and touristic development in the Lake Garda catchment led to escalating pollution from domestic wastewater and runoff. As a result, the lake became more eutrophic with total phosphorous (TP) concentrations increasing by 3-4 times between the 1960s and 2000s. The increase in nutrients led to profound alterations in the lake's planktonic community which may have negatively impacted the food resources available to the Garda Trout. A susequent slight decrease in TP levels may be related to investment in wastewater treatment facilities and other policy-led measures since the 1990s, but this has not yet been confirmed by experimental data. The lake is currently visited by c. 20 million tourists per year, and pollution is likely to increase during the peak summer season.

Lake Garda is undergoing continuous warming due to climate change, and during 2015-2016 the water temperature at depths below 100 metres exceeded all previous records. This trend is also understood to be decreasing the frequency of full mixing events while maintaining relatively high TP levels in the hypolimnion. The effect of these environmental shifts on the Garda Trout has not been investigated, but they could plausibly interfere with its foraging behaviour and reproductive processes. Moreover, the combination of rising water temperatures and longer meromictic periods is driving gradual oxygen depletion in this species' profundal habitat.

Lake Garda supports a commercial fishery, although the number of licensed fishers has fallen from c. 500 to c. 100 since the turn of the 20^th century.

The Garda Trout has been harvested for human consumption since at least the 1500s, particularly during its two reproductive periods when nuptial individuals aggregate at known sites (see 'Habitat and Ecology'). It was traditionally landed using two types of specialised deepwater net known as "réet" and "reù", and later a long-line technique called "tirlindana", but by the mid-20^th century these methods had largely been replaced by pelagic trawl nets towed from two boats and referred to as "volante".

A significant decline in commercial landings after the 1950s (see 'Population') led to the introduction of fisheries regulations during the 1980s. These included closure of the fishery during the reproductive periods, a minimum harvest size of 30 cm, and rules governing the mesh size of trawl nets.

The Garda Trout fishery is currently closed on an indefinite basis, after a total harvesting ban was imposed by the bordering provinces of Brescia and Verona in 2015, followed by Trento in 2019.

Since 2014, a few hundred kilogrammes per year are produced exclusively for human consumption in at least two dedicated fish farms. Concerns have been raised over the use of offspring from these projects for experimental restocking in the lake itself (see 'Conservation').

The southern tip of Lake Garda is included in the European Union's Natura 2000 network (site IT3210018), but it is not understood to cover any of this species' preferred habitat. It was assessed as Critically Endangered under Criterion A (A2bde) for the 2022 iteration of the Italian Red List of Vertebrates, but the data used could not be obtained for the present global assessment.

Efforts to reinforce the wild population through supportive rearing and restocking took place from the 1880s until the 1970s, at which point they were suspended due to concerns over their effectiveness. Each year, gametes were stripped from nuptial individuals by authorised fishers and used for artificial insemination in hatchery facilities, and up to 2,500,000 alevins were released into the lake. However, no correlation between the number of individuals stocked and subsequent commercial yields was ever observed.

More recently, a series of initiatives aiming to reverse the Garda Trout's decline have been implemented by the provinces, e.g. "Carpiogarda" (Trento, since 2008) and "Salvacarpio" (Brescia/Lombardy, since 2011). These activities have largely focused on experimental reproduction and rearing programmes and have resulted in limited restocking, including the trial use of Vibert boxes containing eyed eggs. Captive reproduction has been relatively unsuccesful beyond the maintenance and stripping of broodstocks, and these difficulties have been attributed to the species' somewhat specialised habitat requirements and life history (see 'Habitat and Ecology'). Broodstocks are currently held at a small number of private facilities in each of the three provinces.

Recent research investigating the development of a specific feed to optimise the growth of juveniles in Garda Trout aquaculture has served to highlight ongoing concerns of an entrenched conflict between commercial and conservation objectives within some of the current captive-breeding schemes. In particular, a focus on increasing productivity and the repeated use of captive, semi-domesticated broodstocks for the production of gametes may be creating artificial selection processes or even unintentional hybrids with non-native Brown Trout (Salmo trutta). These approaches are likely to drive genetic drift, accelerate introgression through the release of hybrids, and lead to a loss of hereditary behaviours such as spawning site fidelity. The continued release of such individuals into the lake could thus actively threaten the integrity of the wild gene pool. Moreover, contrasting approaches have been applied in the three provinces, and no area-wide management plan is in place.

Misgivings have also been raised due to a lack of genetic screening and scientific data underpinning some of the current projects, compounded by uncertainty relating to certain aspects of this species' biology and the drivers of its decline.

The use of high-resolution genotyping to identify potential broodstock and strengthen supportive breeding activities is thus strongly recommended. Since the intensive stocking of alevins during the 19^th and 20^th centures failed to improve recruitment over a considerable time period, the introduction of eyed eggs at known spawning sites may represent a promising alternative and should be investigated in depth. Although initial surveys were carried out along a 24.5 km transect in 2010, the quality of existing spawning sites is unclear and should also be explored in order to assess the need for restorative actions.

A deeper understanding of this species' present abundance, population dynamics, life history and response to the identified threats is also required in order to develop a structured management plan, which should ideally be designed and implemented at the cross-provincial scale.