Changes in swimming depths of Atlantic salmon Salmo salar post‐smolts relative to light intensity

Eight hatchery‐reared Atlantic salmon Salmo salar post‐smolts, implanted with acoustic depth sensing transmitters and manually tracked for 5–12 h in the Hardangerfjord (Norway), spent most of their time (49–99%) at 1–3 m depth during the day, whereas four of seven fish tracked were found close (<0·5 m) to the surface at night, with a strong negative cross‐correlation between general swimming depth and surface light intensity. Hence, the actual swimming depth of post‐smolts during their early marine migration may depend on the light conditions, although the individual variation in vertical movement pattern was large. No cross‐correlations were found between light intensity and swimming depth during daytime periods with rapid changes in light intensity, indicating that other factors than light intensity were important in initiating the irregular dives that were recorded down to 6·5 m depth.     

Radiocaesium concentrations in the muscle and eggs of salmonids from Lake Chuzenji,Japan, after the Fukushima fallout

Approximately 18 months (September to December 2012) after the Fukushima Dai‐ichi Nuclear Power Plant accident, elevated radiocaesium concentrations were measured in samples of muscle and eggs from masu salmon Oncorhynchus masou, kokanee Oncorhynchus nerka, brown trout Salmo trutta and lake trout Salvelinus namaycush from the Lake Chuzenji system, central Honshu Island, Japan (160 km from the station). Mean muscle concentrations were 142·9–249·2 Bq kg^?1 wet mass and mean concentrations in eggs were 38·7–79·0 Bq kg^?1 wet mass. There was no relationship between fork length and muscle radiocaesium concentration in any of the species, but there were significant relationships between individual muscle and egg radiocaesium concentrations from O. masou, S. trutta and S. namaycush.     

Keeping close to the river,shore and surface: the first marine migration of brown trout (Salmo trutta) and Arctic charr (Salvelinus alpinus) post-smolts

Acoustic telemetry was utilized to track 49 brown trout (Salmo trutta) and 37 Arctic charr (Salvelinus alpinus) first-time migrants of wild origin [post-smolts; mean L_F (fork length): 169 and 172 mm] in a large fjord in northern Norway. The S. trutta were registered at sea for more than twice the time of the S. alpinus (medians of 54 and 22 days, respectively). Both species were mostly detected near river mouths (>80% of detections) and almost exclusively spent their time (>95%) within the interior 18 km of the fjord. They were surface oriented, with most detections at <1 m depth and S. trutta deeper on average (median mean depths of 0.7 and 0.5 m, respectively). This study concludes that post-smolts of both species stay closer to the surface and to river mouths than larger veteran migrants. This study emphasizes the importance of river mouths and upper water layers for the survival of both species during their first marine migration.     

Summer stream habitat partitioning by sympatric Arctic charr, Atlantic salmon and brown trout in two sub-arctic rivers

Summer habitat use by sympatric Arctic charr Salvelinus alpinus, young Atlantic salmon Salmo salar and brown trout Salmo trutta was studied by two methods, direct underwater observation and electrofishing, across a range of habitats in two sub-arctic rivers. More Arctic charr and fewer Atlantic salmon parr were observed by electrofishing in comparison to direct underwater observation, perhaps suggesting a more cryptic behaviour by Arctic charr. The three species segregated in habitat use. Arctic charr, as found by direct underwater observation, most frequently used slow (mean ±s .d . water velocity 7·2 ± 16·6 cm s⁻¹) or often stillwater and deep habitats (mean ±s .d . depth 170·1 ± 72·1 cm). The most frequently used mesohabitat type was a pool. Young Atlantic salmon favoured the faster flowing areas (mean ±s .d . water velocity 44·0 ± 16·8 cm s⁻¹ and depth 57·1 ± 19·0 cm), while brown trout occupied intermediate habitats (mean ±s .d . water velocity 33·1 ± 18·6 cm s⁻¹ and depth 50·2 ± 18·0 cm). Niche overlap was considerable. The Arctic charr observed were on average larger (total length) than Atlantic salmon and brown trout (mean ±s .d . 21·9 ± 8·0, 10·2 ± 3·1 and 13·4 ± 4·5 cm). Similar habitat segregation between Atlantic salmon and brown trout was found by electrofishing, but more fishes were observed in shallower habitats. Electrofishing suggested that Arctic charr occupied habitats similar to brown trout. These results, however, are biased because electrofishing was inefficient in the slow-deep habitat favoured by Arctic charr. Habitat use changed between day and night in a similar way for all three species. At night, fishes held positions closer to the bottom than in the day and were more often observed in shallower stream areas mostly with lower water velocities and finer substrata. The observed habitat segregation is probably the result of interference competition, but the influence of innate selective differences needs more study.     

Latitudinal and altitudinal growth patterns of brown trout Salmo trutta at different spatial scales

Spatial variation in growth of stream‐dwelling brown trout Salmo trutta was explored in 13 populations using a long‐term study (1993–2004) in the Bay of Biscay drainage, northern Spain. The high variability in fork length (L_F) of S. trutta in the study area was similar to the body‐size range found in the entire European distribution of the species. Mean L_F at age varied: 0+ years, 57·4–100·7 mm; 1+ years, 111·6–176·0 mm; 2+ years, 155·6–248·4 mm and 3+ years, 194·3–290·9 mm. Average L_F at age was higher in main courses and lower reaches compared with small tributaries and upper reaches. Annual specific growth rates (G_L) were: 0+ to 1+ years, 0·634–0·825 mm mm⁻¹ year⁻¹; 1+ to 2+ years, 0·243–0·342 mm mm⁻¹ year⁻¹; 2+ to 3+ years, 0·166–0·222 mm mm⁻¹ year⁻¹, showing a great homogeneity. Regression models showed that water temperature and altitude were the major determinants of L_F at age variability within the study area. A broader spatial analysis using available data from stream‐dwelling S. trutta populations throughout Europe indicated a negative relationship between latitude and L_F of individuals and a negative interaction between latitude and altitude. These findings support previous evidence of the pervasive role of water temperature on the L_F of this species. Altitude appeared as the overall factor that includes the local variation of other variables, such as water temperature or food availability. At a larger scale, latitude was the factor that encompassed these environmental gradients and explained the differences in L_F of S. trutta. In summary, L_F at age in stream‐dwelling S. trutta decreases with latitude in Europe, the converse of Bergmann's rule.     

The first months at sea: marine migration and habitat use of sea trout Salmo trutta post‐smolts

The early migration and habitat use of brown trout Salmo trutta post‐smolts tagged with acoustic transmitters (n = 50) were investigated in a fjord system in central Norway from 30 April to 26 November 2014. The main aims were to investigate return rate, marine residence time and spatial use of the fjord system. Median seaward migration and return to fresh water dates were 22 May and 4 July, respectively. Of the 40 seaward migrating smolts, 26 returned to fresh water, giving a minimum return rate to fresh water of 65%. Entrance to the fjord from the river occurred mainly at night (80% of the S. trutta), however, no such diurnal pattern was observed during the return migration. Mean marine residence time was 38 days, but with large individual variation (22–99 days). The innermost parts of the study area were more utilized than the outer part of the fjord system during the sea residency, and with more use of the near shore habitat than the open, pelagic areas. Many post‐smolts also utilized the outer part of the fjord system, however, and 94% of the post‐smolts were recorded at least 14 km from the home river mouth. Marine survival and distribution in the fjord were size dependent with the largest individuals utilizing outer fjord areas and having higher return rates to fresh water. As far as is known, this is the first published study on temporal and spatial behaviour in the marine environment of first‐time S. trutta migrants during the full course of their first trip to sea.     

Vertical movement of dolphinfish <Emphasis Type="Italic">Coryphaena hippurus</Emphasis> as recorded by acceleration data-loggers in the northern East China Sea

Environmental changes influence foraging behavior for most animals. Dolphinfish, Coryphaena hippurus, are epipelagic predators and have a cosmopolitan tropical to warm-temperate (>20°C) distribution. We simultaneously obtained the ambient temperature and the foraging behavior (i.e., swimming speed, depth and tailbeat acceleration) of dolphinfish, using an acceleration data-logger in May, September, October, November 2007, June 2008, May and July 2010 for 8 individuals. Although the dolphinfish spent a mean ± standard deviation of 43.4 ± 27.7% of their time at the surface (0–5 m), dive excursions from the surface (DES) were observed in all individuals and maximum DES depths ranged from 50.1 to 95.4 m. DES events resulted dives below the thermocline for these dolphinfish, and there was a significantly positive relationship between the isothermal layer depth (ILD) and DES depth. Our results demonstrate that dolphinfish avoided the rapid thermal change beyond the thermocline, and their prey is most likely found in the upper layers of the thermocline. Gliding behavior during the DES phase was also observed and dolphinfish gradually descended to deeper waters with gliding. The gliding time was longer when the ILD was deeper, and fish tended to dive deeper. We suggest that dolphinfish adopt gliding behavior to search a broader range of depths for prey, while minimizing energy use.     

Genetic assessment of Atlantic salmon Salmo salar and sea trout Salmo trutta stocking in a Baltic Sea river

Microsatellite DNA variation was used to assess the outcome of stocking Atlantic salmon Salmo salar and migratory trout Salmo trutta in River Sävarå, N Sweden. No information on pre‐stocking genetic composition of S. salar and S. trutta in River Sävarå was available. In 2 year‐classes of S. salar smolt, microsatellite data indicated that post‐stocking genetic composition differed markedly (F_ST= 0·048) from the main donor strain, Byskeälven S. salar, and from other Gulf of Bothnia S. salar stocks (F_ST 0·047 and 0·132). The STRUCTURE programme failed to detect any substructuring within Sävarå salmon. It was concluded that only minor introgression estimated to a proportion of 0·11 (95% CI 0·07–0·16) has occurred in S. salar. Salmo trutta showed overall low differentiation among populations with maximum F_ST of 0·03 making analysis more cumbersome than in S. salar. Still, the SävaråS. trutta deviated significantly from potential donor populations, and STRUCTURE software supported that majority of trout in Sävarå formed a distinct genetic population. Admixture was more extensive in S. trutta and estimated to 0·17 (95% CI 0·10–0·25).     

The impact of intra- and interspecific interactions on young-of-the-year brook charr, in temperate lakes

The abundance, growth, spatial distribution, and feeding habits of five allopatric brook charr, Salvelinus fontinalis, populations (young-of-the-year, 0+ juveniles; YOY) were compared with five other populations living sympatrically with white sucker, Catostomus commersoni. The study was made in oligotrophic lakes of the Laurentian Shield (Québec, Canada) during three sampling periods in 1989 (July, August and September). The abundance of YOY charr was significantly higher in allopatric than in sympatric populations (45·3 ± 3·8 vs 3·4 ± 3·8 fish/lake caught in 1773 m² of gillnets; P<0·005). The mean length of YOY charr did not differ among allopatric and sympatric populations at each sampling period; July: 60·2 ± 3·0 vs 60·0 ± 4·5 mm; August: 61·9 ± 4·5 vs 63·2 ± 4·1 mm; September: 77·9 ± 8·7 vs 77·3 ± 7·8 mm respectively. Horizontal distribution of allopatric YOY charr did not differ from that of sympatric charr, 65% of the fish being captured within the first 2 m depth and the rest between 2 and 7 m depth. In contrast, the vertical distribution of allopatric YOY charr from both communities was significantly different; 81% of allopatric charr were captured within 0·5 m from the substrate compared to 64% for sympatric charr (P<0·001). Differences in vertical distribution of the fish were related to differences in diet; allopatric charr fed mainly on benthic and large planktonic organisms whereas sympatric charr fed less on these organisms and more on terrestrial organisms. In the lake where YOY charr were most abundant, individuals were spatially segregated into two groups; one ‘littoral’, found in 0–2m depth, and one ‘profundal’, found in 3–6 m depth. Growth, condition, and feeding habits of charr from the two groups were different, especially during the last sampling period.     

pH preference and avoidance responses of adult brook trout Salvelinus fontinalis and brown trout Salmo trutta

The pH preferred and avoided by wild, adult brook trout Salvelinus fontinalis and brown trout Salmo trutta was examined in a series a laboratory tests using gradual and steep‐gradient flow‐through aquaria. The results were compared with those published for the observed segregation patterns of juvenile S. fontinalis and S. trutta in Pennsylvania streams. The adult S. trutta tested showed a preference for pH 4·0 while adult S. fontinalis did not prefer any pH within the range tested. Salmo trutta are not found in Pennsylvania streams with a base‐flow pH < 5·8 which suggests that S. trutta prefer pH well above 4·0. Adult S. trutta displayed a lack of avoidance at pH below 5·0, as also reported earlier for juveniles. The avoidance pH of wild, adult S. fontinalis (between pH 5·5 and 6·0) and S. trutta (between pH 6·5 and 7·0) did not differ appreciably from earlier study results for the avoidance pH of juvenile S. fontinalis and S. trutta. A comparison of c.i. around these avoidance estimates indicates that avoidance pH is similar among adult S. fontinalis and S. trutta in this study. The limited overlap of c.i. for avoidance pH values for the two species, however, suggests that some S. trutta will display avoidance at a higher pH when S. fontinalis will not. The results of this study indicate that segregation patterns of adult S. fontinalis and S. trutta in Pennsylvania streams could be related to pH and that competition with S. trutta could be mediating the occurrence of S. fontinalis at some pH levels.     

Spatial ecology of coastal Atlantic cod Gadus morhua associated with parasite load

Acoustic tags and receivers were used to investigate the spatial ecology of coastal Atlantic cod Gadus morhua (n = 32, mean fork length: 50 cm, range: 33–80 cm) on the Norwegian Skagerrak coast in 2012. Monthly home ranges (HR), swimming activity and depth use varied considerably among individuals and through the months of June, July and August. HR sizes for the period ranged from 0·25 to 5·20 km² (mean = 2·30 km²). Two thirds of the tagged G. morhua were infected with black spot disease Cryptocotyle lingua parasites; these fish had larger HRs and occupied deeper water compared with non‐infected fish. The infected fish also tended to be more active in terms of horizontal swimming. From an ecological and evolutionary perspective, any environmental change that modifies G. morhua behaviour may therefore also alter the parasite load of the population, and its conservation and fishery status.     

Habitat use and early winter movements by juvenile Atlantic cod in a coastal area of Newfoundland

Acoustic telemetry was used to monitor the movements and landscape scale habitat use of age 2–3 year juvenile Atlantic cod Gadus morhua in Newman Sound, a coastal fjord of Newfoundland, during late autumn and early winter (24 October 1999 to 22 January 2000). Substratum, bathymetric relief and depth use were studied to determine if it differed from the pattern expected given an absence of selectivity (habitat use proportional to habitat availability). Prior to winter migrations, most Atlantic cod maintained small home ranges (0·5–33·4 ha, median = 2·1 ha) although a few individuals moved more widely. No relationship between total length (L_T) and home range size was detected. In Inner Newman Sound, Atlantic cod occupied depths of 10–29 m more than expected given availability, while depths of 0–9 and 40–59 m were underutilized. No significant relationship between depth and L_T was detected. Areas of medium (5–10%) or high (>10%) bathymetric relief and boulder or kelp habitats were used significantly more than expected given the availability of these habitats. Sand and eelgrass substrata were underutilized given availability, although many Atlantic cod used sandy‐bottomed areas to some degree. Flexibility in habitat use by the individuals that were studied suggested reduced predation risk relative to younger conspecifics. Winter migrations to deeper water beyond Newman Sound began in mid‐November, coinciding with the disappearance of the thermocline, and continued until 27 December. Approximately 30% of monitored individuals did not migrate and maintained their home ranges into the winter season. No significant differences in L_T between migrating and resident groups were detected, however, the condition of migrating fish was significantly higher than in resident fish. This finding supports the theory that feeding history plays a role in the decision to undertake migration.     

Static habitat partitioning and dynamic selection by sympatric young Atlantic salmon and brown trout in south-west England streams

Direct underwater observation of micro‐habitat use by 1838 young Atlantic salmon Salmo salar [mean L_T 7·9 ± 3.1(s.d.) cm, range 3·19] and 1227 brown trout Salmo trutta (L_T 10·9 ± 5·0 cm, range 3·56) showed both species were selective in habitat use, with differences between species and fish size. Atlantic salmon and brown trout selected relatively narrow ranges for the two micro‐habitat variables snout water velocity and height above bottom, but with differences between size‐classes. The smaller fishes <7 cm held positions in slower water closer to the bottom. On a larger scale, the Atlantic salmon more often used shallower stream areas, compared with brown trout. The larger parr preferred the deeper stream areas. Atlantic salmon used higher and slightly more variable mean water velocities than brown trout. Substrata used by the two species were similar. Finer substrata, although variable, were selected at the snout position, and differences were pronounced between size‐classes. On a meso‐habitat scale, brown trout were more frequently observed in slow pool‐glide habitats, while young Atlantic salmon favoured the faster high‐gradient meso‐habitats. Small juveniles <7 cm of both species were observed most frequently in riffle‐chute habitats. Atlantic salmon and brown trout segregated with respect to use of habitat, but considerable niche overlap between species indicated competitive interactions. In particular, for small fishes <7 cm of the two species, there was almost complete niche overlap for use of water depth, while they segregated with respect to water velocity. Habitat suitability indices developed for both species for mean water velocity and water depth, tended to have their optimum at lower values compared with previous studies in larger streams, with Atlantic salmon parr in the small streams occupying the same habitat as favoured by brown trout in larger streams. The data indicate both species may be flexible in their habitat selection depending on habitat availability. Species‐specific habitat overlap between streams may be complete. However, between‐species habitat partitioning remains similar.     

Differences in growth between offspring of anadromous and freshwater brown trout Salmo trutta

In this study, individual growth of juvenile offspring of anadromous and freshwater resident brown trout Salmo trutta and crosses between the two from the River Imsa, Norway, was estimated. The juveniles were incubated until hatching at two temperatures (±S.D. ), either 4.4 ± 1.5°C or 7.1 ± 0.6°C. Growth rate was estimated for 22 days in August–September when the fish on average were c. 8 g in wet mass, and the estimates were standardized to 1 g fish dry mass. Offspring of anadromous S. trutta grew better at both 15 and 18°C than offspring of freshwater resident S. trutta or offspring of crosses between the two S. trutta types. This difference appears not to result from a maternal effect because anadromous S. trutta grew better than the hybrids with anadromous mothers. Instead, this appears to be an inherited difference between the anadromous and the freshwater resident fish lending support to the hypothesis that anadromous and freshwater resident S. trutta in this river differ in genetic expression. Egg incubation temperature of S. trutta appeared not to influence the later growth as reported earlier from the studies of Atlantic salmon Salmo salar.     

Mitigation of acidified salmon rivers – effects of liming on young brown trout Salmo trutta

In southern Norway, 22 acidified rivers supporting anadromous salmonids were mitigated with lime to improve water quality and restore fish populations. In 13 of these rivers, effects on Salmo trutta and Salmo salar densities were monitored over 10–12 years, grouped into age 0 and age ≥ 1 year fish. These rivers had a mean annual discharge of between 4·9 and 85·5 m³ s^?1, and six of them were regulated for hydro‐power production. Salmo salar were lost in six of these rivers prior to liming, and highly reduced in the remaining seven rivers. Post‐liming, S. salar became re‐established in all six rivers with lost populations, and recovered in the seven other rivers. Salmo trutta occurred in all 13 study rivers prior to liming. Despite the improved water quality, both age 0 and age ≥ 1 year S. trutta densities decreased as S. salar density increased, with an average reduction of >50% after 10 years of liming. For age 0 year S. trutta this effect was less strong in rivers where S. salar were present prior to liming. In contrast, densities of S. trutta increased in unlimed streams above the anadromous stretches in two of the rivers following improved water quality due to natural recovery. Density increases of both age 0 and age ≥ 1 year S. salar showed a positive effect of river discharge. The results suggest that the decline in S. trutta density after liming is related to interspecific resource competition due to the recovery of S. salar. Thus, improved water quality through liming may not only sustain susceptible species, but can have a negative effect on species that are more tolerant prior to the treatment, such as S. trutta.     

Dietary ontogeny and niche shift to piscivory in lacustrine brown trout Salmo trutta revealed by stomach content and stable isotope analyses

The feeding ecology and ontogeny of a large size range of brown trout Salmo trutta in Lake Fyresvatnet, southern Norway, were examined by stomach content and stable isotope analyses. According to the stomach contents, the S. trutta changed their diet at c. 30 cm total length (L_T). The smaller size classes fed on benthic invertebrates and surface insects, whereas larger S. trutta (>30 cm) fed mainly on whitefish Coregonus lavaretus. A similar, but more gradual shift to piscivory in the size range 25–30 cm was found when using the stable isotope mixing model SIAR to reveal dietary ontogeny. The δ¹³C isotopic signature confirmed that S. trutta independent of size predominantly relied upon benthic energy sources, suggesting that the littoral zone was the primary foraging habitat for both invertebrate and piscivorous feeders. The δ¹⁵N values and trophic position increased with predator length, ranging from an average of 3·60 for small‐sized S. trutta (<15 cm) to 4·15 for large‐sized fish (>35 cm). The S. trutta exhibited a relatively slow growth rate during the predominant invertebrate feeding stages up to 7 years of age and 28 cm L_T, whereas fish above this size and age displayed a rapid growth rate of 9–11 cm year^?1, demonstrating the profitability of piscivorous feeding.     

Atlantic sturgeon Acipenser oxyrinchus Mitchill, 1815 seasonal marine depth and temperature occupancy and movement in the Bay of Fundy

Pop‐up satellite archival tags were used to collect fisheries‐independent data that characterized the seasonal habitat occupancy and movement of Atlantic sturgeon Acipenser oxyrinchus Mitchill, 1815 in the Bay of Fundy (BoF). Atlantic sturgeon from Canadian and United States stocks aggregate annually for feeding in Minas Basin, inner BoF (45.28N, 64.18W), during May to September but depart to other locations from October to April. Sixteen PSAT tags were applied to sturgeon ranging from 152 to 203 cm total length captured and released in Minas Basin during May to August. Ten of the tags were reported after 1 year at large and provided pop‐off locations. Seven were recovered with archived data, or provided transmitted data sets, which were analyzed for depth and temperature occupancy from June 2011 to August 2013. During June to August while in the Minas Basin the sturgeon spent >90% of time at depths <10 m and temperatures of 16–22°C. Departure from Minas Basin through the Minas Passage was in September and October, when depth occupancy varied from <10 to 120 m. From November to April sturgeon were in the outer BoF, where mean depth occupancies ranged from 40 to 100 m at mean temperatures of 3–14°C. Deepest mean depth occupancy of 60 to 90 m was recorded during December 2011 and 2012, and coldest mean temperature occupancy of 0–4°C in March 2011 and 2012. During April and May mean depth and temperature occupancy ranges shallowed from 40 to <10 m and increased from 4 to 15°C, respectively. Tag pop‐off locations indicated that sturgeon spent the winter season in the outer BoF but by June had either migrated back to the Minas Basin or off the mouth of the Saint John River, a known spawning location.     

Riverine and near coastal migration performance of hatchery brown trout Salmo trutta

To study migration performance and return rates of hatchery brown trout Salmo trutta smolts the first 5 months after release, 50 fish in each year (fork length, L_F, 158–288 mm) were in two subsequent years tagged with acoustic transmitters and recorded by automatic listening stations in the River Nidelva (central Norway), its estuary and in the marine environment. More than half of the smolts became anadromous migrants (52% in 2011 and 70% in 2012). The fish spent longer time in the estuary than in the marine environment and the results suggest that migratory behaviour of S. trutta smolts is not only restricted to be resident or anadrome–lacustrine, but that there is also an intermediary strategy of estuarine feeding. There were no differences in L_F or mass between groups of smolts with different migration patterns. Return rates from the sea within the first 5 months after release were in both years 16%. Median progression rate in the river was 0·090 L_F s^?1 but decreased significantly as the smolts entered the estuary (0·015 L_F s^?1). The long residential time in the estuary may increase the risk of negative effects of anthropogenic activities in estuaries, such as harbours and industrial development, and special attention should be given to evaluate effects of such activities.     

Colonization of the deep sea by fishes

Analysis of maximum depth of occurrence of 11 952 marine fish species shows a global decrease in species number (N) with depth (x; m): log₁₀N = ?0·000422x + 3·610000 (r² = 0·948). The rate of decrease is close to global estimates for change in pelagic and benthic biomass with depth (?0·000430), indicating that species richness of fishes may be limited by food energy availability in the deep sea. The slopes for the Classes Myxini (?0·000488) and Actinopterygii (?0·000413) follow this trend but Chondrichthyes decrease more rapidly (?0·000731) implying deficiency in ability to colonize the deep sea. Maximum depths attained are 2743, 4156 and 8370 m for Myxini, Chondrichthyes and Actinopterygii, respectively. Endemic species occur in abundance at 7–7800 m depth in hadal trenches but appear to be absent from the deepest parts of the oceans, >9000 m deep. There have been six global oceanic anoxic events (OAE) since the origin of the major fish taxa in the Devonian c. 400 million years ago (mya ). Colonization of the deep sea has taken place largely since the most recent OAE in the Cretaceous 94 mya when the Atlantic Ocean opened up. Patterns of global oceanic circulation oxygenating the deep ocean basins became established coinciding with a period of teleost diversification and appearance of the Acanthopterygii. Within the Actinopterygii, there is a trend for greater invasion of the deep sea by the lower taxa in accordance with the Andriashev paradigm. Here, 31 deep‐sea families of Actinopterygii were identified with mean maximum depth >1000 m and with >10 species. Those with most of their constituent species living shallower than 1000 m are proposed as invasive, with extinctions in the deep being continuously balanced by export of species from shallow seas. Specialized families with most species deeper than 1000 m are termed deep‐sea endemics in this study; these appear to persist in the deep by virtue of global distribution enabling recovery from regional extinctions. Deep‐sea invasive families such as Ophidiidae and Liparidae make the greatest contribution to fish fauna at depths >6000 m.