Early ontogeny of the lesser sandeel (Ammodytes marinus)

Lesser sandeel (Ammodytes marinus) is widely distributed in North Sea ecosystems. Sandeel acts as a critical trophic link between zooplankton and top predators (fish, mammals, sea birds). Because they live buried in the sand, sandeel may be directly affected by the rapid expansion of anthropogenic activities linked to their habitat on the sea bottom (e.g., hydrocarbon extraction, offshore renewable energy, and subsea mining). It is, therefore, important to understand the impact of cumulative environmental and anthropogenic stressors on this species. A detailed description of the ontogenetic timeline and developmental staging for this species is lacking limiting the possibilities for comparative developmental studies assessing, e.g., the impact of various environmental stressors.


| INTRODUCTION
Lesser sandeel (Ammodytes marinus [Raitt, 1934]) are very abundant and widely distributed in the North Atlantic and in the southwestern Baltic Sea. 1 In the North Sea, sandeel represents up to 15% of the total fish biomass and it is estimated that 90% of it is lesser sandeel (hereafter, sandeel). 2[7][8][9][10][11] Changes in the abundance of sandeel have impacts on the food chain, from larger fish to seabirds, which display dramatically reduced breeding success when local sandeel stocks decrease. 12andeel juveniles and adults spend much of their time buried in oxygen-rich sand or gravel sediments with specific granulometry at depths from 10 to 150 m. 4,13,14 This lifestyle conserves energy and provides protection from predators. 10,157][18][19][20][21] During the feeding season, a large number of individuals emerge from the seabed in large pelagic schools to prey upon zooplankton.These schools form bridging structures that likely play an important role in their ecology and prevent post-settled juveniles from being disconnected from suitable bottom substrate. 225][26] After a long (1 month) embryonic period, larvae hatch and drift with the current in the deep layer.After metamorphosis, juveniles congregate and settle in the same areas where the adults are found. 27Although they form large schools, post-settled juvenile and adult sandeel display high site fidelity. 28,29uring the winter (the spawning period), sandeels rarely emerge from the seabed to feed, subsisting on stored energy.
Considering the ecological and commercial importance of sandeel, risk assessments of the cumulative environmental and anthropogenic stressors impacting them are needed.Due to the rapidly expanding habitatchanging man-made structures and activity linked to the sea bottom, such as offshore renewable energy, oil and natural gas extraction platforms, and subsea mining, knowledge of the biology of sandeel is needed.Assessing any anthropogenic or climate-driven impacts on sandeel requires a better understanding of the sensitivity of their early life stages.To conduct any such study, it is essential to have a detailed table of all the developmental features and rearing methods that provide guidance for assessing fertility, hatchability, and larval survival and behavior.
Although the main life stages in some sandeel species (A.americanus, A. personatus) have been described, a detailed systematic description of the morphological development of lesser sandeel is unavailable.][32][33] The objective of this study is to provide a detailed description of the embryonic and larval biology of lesser sandeel and to outline their developmental trajectory.This study also describes methods for gamete stripping and intensive culture techniques.

| Spawning evaluation and comparison of fertilization substrates
Gamete (sperm and eggs) collection was successfully performed by hand-stripping the males and females.Excessive pressure applied along the abdomen was previously reported to rupture membranes holding the eggs, injuring the ovaries, and resulting in lower egg survival. 31In this study, gentle pressure along the abdomen was sufficient to assist the natural flow of eggs.However, stripping often resulted in the death of the female during the days that followed.Therefore, we advise that females be euthanized after stripping.The weight of spawned eggs represented 16% of the female total weight (Figure 1A).
Egg fertilization was performed on different substrates (glass slide, filter mesh [500, 300, and 100 μm], and sand substrate).A comparison of substrates is summarized in Table 1.The fertilization rates and viability of embryos (48 hours post-fertilization, hpf) were 70%-80% for the egg batches placed in tubes on filter mesh regardless of mesh pore size (100, 300, and 500 μm).Fertilization from egg batches placed on microscope slides varied from 40% to 80%, whereas no fertilized eggs were found in the batches placed on natural sand.Fertilization and incubating sandeel embryos in natural sand was not a successful option, leading to a high number of inconveniences for experimental studies (no observations possible without disturbance, no control of egg quality and survival).This result was surprising since it is the natural egg substrate; the reason for it is unclear although possibilities include insufficient oxygenation of the sand or too rough handling of eggs.The glass microscope slide substrate was the most appropriate method for observing embryo development and sampling while introducing minimal disturbance to the embryos.Observations could be made on every slide and the whole egg batch could be observed with minimal disturbance.Embryos could be efficiently sampled and dying individuals can be removed using this system.However, bacterial growth was observed on the slides, and hatching success was poor resulting in very few surviving larvae (<30%-40%).The hatching period on glass slides extended to over a month and many of the embryos did not hatch before they were mechanically stimulated with a flow of seawater.The prolonged embryonic period resulted in smaller larvae and higher mortality in hatched larvae and in unhatched embryos.The development of embryos and larvae from the glass slides also varied according to the number of egg layers on the slide (Figure 1B).Areas with more than three clustered embryos showed slower development compared to single embryos or embryos in smaller groups (Figure 1B), suggesting that the clustered eggs did not receive enough oxygenation or water flow to grow properly and were possibly also more susceptible to bacterial infection.The eggs placed in tubes on the filter mesh were more difficult to observe without physically removing them.Moving the filter system increased the chance of detaching the eggs from the substrate (due to water flow through the filter), increasing the chance of damaging and/or losing eggs in the rearing tank.The 500 μm filter mesh was too big, leading to the compression of some eggs trapped in the mesh opening.While the filter mesh systems (500, 300, and 100 μm) were more inconvenient for embryonic observations, embryos received more oxygenation and water flow compared to the batches placed on the glass slides, limiting the bacterial growth, and resulting in egg development being more consistent and synchronized (observations made from sampled embryos under the microscope).The hatching success was also high in the filter mesh systems; >70%-80% hatching and a shorter overall period for the batch to hatch.Hatching duration in the filter mesh systems was 8-9 days with a peak at 3 days after the first hatching had begun compared to over a month for the glass slide system.Synchronized hatching required good oxygenation and stimulation with sufficient water flow going over the eggs.Adult sandeel require well-flushed, tidally active areas that are oxygen-rich to survive and spawn. 4,13,14,34,35Embryogenesis and survival of early life stages are influenced by temperature and oxygen concentration.Low temperature and oxygen concentration delay egg development in sandeel. 11,30,31The literature on the embryonic development of sandeel is limited.Smigielski, Halavik, Buckley, Drew, and Laurence 31 reported an incubation and hatch duration for A. americanus of between 55 and 135 days at 10 C and 2 C, respectively.The authors suggested a lengthy hatching period for this species, although it is still uncertain if the hatching duration is as long in the natural environment.The rearing systems using a filter mesh size of 300 μm produced an outcome consistent with the hatching duration reported by Régnier, Gibb, and Wright 30 for A. marinus.

| Early development
Embryonic and larval development in sandeel has only been partially described, 33 making it difficult to compare our results with those reported in earlier studies.From a developmental perspective, sandeel share numerous physiological and morphological features with many teleost fishes, particularly herring. 31,36,37The most striking feature specific to sandeel and herring species is the long embryonic development (a month) with a long elongated pharyngula.A distinctive character in sandeel larvae is also the position of the anus; the ratio of pre-anal and post-anal body length is 4:3, compared 4:1 in herring.
Sandeel eggs are almost spherical with a diameter ranging from 0.85 to 1.23 mm (mean ± SD; 1.03 ± 0.07 mm; n = 1220 eggs).They have a semitransparent adhesive layer and become brownish after fertilization, making observations of the embryo challenging.The eggs adhered immediately to the substrate after fertilization.Eggs have a clearly visible micropyle and the embryo contains a single oil globule measuring 0.021 ± 0.002 mm 3 (Figure 2A).
The developmental timing and characteristics were aggregated from data acquired at 7.8 C ± 0.5 C, a salinity of 34 ppt under a photoperiod of 12 h light/12 h dark.The developmental timings are expressed in degree days (DD) and normalized as hours or days post-fertilization/ post-hatching (Table 2).
The cleavage, blastula and gastrula periods could be identified based on embryonic morphology.The first cleavage occurs a couple of hours after fertilization and results in two blastomeres of equivalent size, as observed in most teleost fishes.The second cleavage division (Figure 2A) occurs at 1.7 DD and cell division and migration continue thereafter.In sandeel, cleavage is meroblastic; a blastula appeared with a well-formed blastodisc and perivitelline space.The yolk sac syncytial layer is evident and a multilayered and highly domed blastoderm was formed at 10 DD (Figure 2B).The germ ring is then well defined at 15 DD, and the first epiboly movements begin.Gastrulation takes place with the appearance of the embryonic shield (Figure 2C).The cellular front reaches 50% epiboly at 18 DD (Figure 2C), 70% at 26 DD, and covers 100% at 35 DD.The end of the epiboly process indicates the beginning of the segmentation process with the formation of the first somites (38 DD; 10+ somites).The epiboly and segmentation steps were also successive in Pacific herring and in zebrafish. 38,39The somites progressively increase in number, reaching 20-25 somites at 50 DD (Figure 2D-D 0 ) and 54+ somites at 67 DD (Figure 2E-E 0 ).At the same time, organogenesis begins.The optic placode is evident at 50 DD (Figure 2D-D 0 ).The shape of the yolk extension continues to distinguish it from the anterior region.At 60-67 DD (SL = 2.3 ± 0.05 mm), the lens and otic vesicles are well developed, the heart tube is visible, and the end of the tail is detached from the yolk sac (Figures 2E-E  A striking feature specific to sandeel and herring species is the long duration of the pharyngula stage with an elongated tail twisting around the yolk. 36The tail twists around the yolk sac several times and the embryo (Q, Q 0 ) Late flexion larvae (39 dph).R, Juvenile (4 months).S, post-settled juvenile, young adult phase (8 months); bhg, blue hindgut; br, brain; C, cells; DD, degree days; dph, days post-hatching; dt, detached tail from yolk; es, embryonic shield; fi, dorsal and anal fins; fr, fin rays; gr, germ ring; hc, hatching cells; he, heart; ld, lipid droplet; le, lens; m, mouth; me, melanophores; mp, micropyle; op, optic primordium; pec, pectoral fin; pi, pigmentation; pp, posterior pole; so, somite; ys: yolk sac; ysl, yolk syncytial layer.Scale bars A-K, K 00 , L, L 00 , M, M 00 , N, N 00 , O, O 00 , P, P 00 , Q = 100 μm.Scale bars K 0 , L 0 , M 0 , N 0 , O 0 , P 0 , Q 0 = 500 μm.Scale bars R, S = 0.5 cm.Photo credit: P. Perrichon.On the tube mesh substrate, the hatching period lasted for several days (8-9 days; from 207 to 270 DD) with a peak hatching at 3 days from the start of hatching (229 DD).Newly hatched larvae measured 5.93 ± 0.41 mm (Figure 2L-L 00 ), which is consistent with the size reported in the literature (ca.5.5 mm at hatching 33 ).The larvae are transparent, the head is straight along the body axis although the mouth is not yet fully open at hatching.The yolk sac is still present and still contains a lipid droplet.Pectoral fins and gill arches are well formed and movement of the pectoral fins is noted, along with increased swimming.At the peak of hatching (229 DD; Figure 2M-M 00 ), the mouth gape and the jaw expand.Exogenous feeding begins 3 days after hatching when sandeel larvae start to actively hunt prey.Similar to herring species, sandeel larvae may show evidence of enterohepatic recycling for taurine into the hindgut ("blue hindgut") between 2 dph/229 DD and 11-12 dph/310-320 DD (Figure 2M 0 -M 00 , N 0 -N 00 , O 0 -O 00 ).Taurine is crucial in the development of some teleost larvae to maintain the antioxidant defense, redox homeostasis, and oxidative stress necessary for growth and metamorphosis. 40Further analyses should be undertaken to confirm this physiological pathway.The first appearance of xanthophores is seen along the body at 7 dph/270 DD, along with a slight patch of melanophores appearing on the dorso-posterior side of the head and on the inferior part of the caudal fin (Figure 2N-O 00 ).The intensity of pigmentation increases with further development.The yolk sac and lipid droplet were completely resorbed after 11 dph/310 DD.From 22 dph/400 DD, sandeel enters the flexion stage (Figure 2P-Q 0 ).Flexion is defined as follows: (i) pre-flexion (notochord flexion 0-4 ) with up to 10 fin rays is observed in 9.83 ± 0.39 mm larvae, (ii) early flexion (notochord flexion 10-25 ) with greater than or equal to 10 fin rays is observed in 11.20 ± 0.49 mm larvae, and (iii) flexion/post-flexion (notochord flexion 30-45 ) with greater than 13 fin rays is observed in 12.59 ± 0.58 mm larvae (39 dph; Figure 2Q-Q 0 ).Dorsal and anal fins start to develop from the early flexion stage.
At this point, sandeel enters the flexion stage, and the musculature is evident throughout the body and the eyes and mouth are prominent and fully developed.Fin development progresses and displays the adult configuration (Figure 2R, S).The body coloration continues to progress, and the fish reaches 95-100 mg (SL = 3.3 cm, Figure 2R).At 5 months, body coloration is adult-like.At this stage, sand was offered to the fish for settlement.Settlement into sandy habitat occurs at a length of 40-50 mm in the wild. 1,29In our rearing system, fish continued to grow, reaching 7.96 ± 0.90 cm after 8 months (Figure 2S).This body length is similar to our own measurements of wild juveniles, likely the same age, caught during a cruise carried out at the same time.
This study provides the foundation for further comparative developmental studies and a biological basis to better assess the condition of sandeel exposed to environmental and anthropogenic stressors.A detailed description of the developmental stages and key ontogenetic events from fertilization to post-flexion is also provided.Further studies should be undertaken to improve the quality of juvenile stages in intensive culture systems, including assessing the nutritional needs of sandeel larvae.

| Ethics statement
The Austevoll Research Station, Institute of Marine Research, Storebø, Norway is a certified Research Animal Facility for fish of all developmental stages and has permits from the Norwegian Directorate of Fisheries to catch and maintain sandeel.All experimental protocols and procedures were performed in accordance with approved guidelines.

| Wild-caught broodstock
Lesser sandeel (Ammodytes marinus) were collected off the coast of Karmøy, Rogaland, Norway (59.255806N latitude, 5.187222 E longitude) in December 2020 and 2021 using a sandeel dredge deployed from a 10-m long fishing vessel, the Åkrabuen.The individuals were transferred to a 120 L transport tank containing seawater at a salinity of 34 ppt at 8 C. The bottom of the transport tank was filled with sand from the sandeel habitat.The sex ratio of mature fish was three males to one female.
Once at the Austevoll Research Station, fish were transferred to a 1500 L tank.The conditions in the tank matched, as much as possible, those in the natural habitat where the sandeels were collected.A 10-cm layer of sand (0.8-1.25 mm granulometry) covered the bottom, a 12 h light/12 h dark photoperiod was applied and there was a continuous inflow of filtered seawater temperature at 8 C. Fish were not fed until spawning because they do not feed in the wild during the spawning season. 15,25tarting 1 week before their natural spawning period (according to sandeel fishers during the second week of January), the sandeel were examined weekly to assess their reproductive status.

| Gamete collection and fertilization
Mature and ready-to-spawn individuals were selected from the broodstock tank.The average size of mature fish was 14.7 ± 1.4 cm.Maturity could be assessed as males had running milt and females released eggs when the abdomen was pressed gently.Pasteur pipettes were used to collect milt, and eggs were placed in different fertilization substrates prior to fertilization (see section 3.4.for a description of the different substrates).Each female exhibited one unique spawning event during the reproduction season, while the males had running milt throughout the period.As the eggs turn sticky shortly after fertilization, particular care was taken, as much as possible, to ensure that eggs were evenly distributed in no more than 2-3 layers on substrates of different types (see section 2.4.).Fertilization was carried out by adding milt and seawater to the eggs placed on the substrates.The fertilization process lasted 45 minutes before rinsing, after which each substrate with fertilized eggs was transferred to a 50 L tank.

| Substrates for egg development
To establish the best incubation system for egg development, fertilization of the stripped eggs was attempted in five types of substrates (Figure 4): (i) glass microscope slides, 10 L cylindrical tubes fitted with filter mesh with pores sizes of (ii) 500 μm, (iii) 300 μm, (iv) 100 μm, and (v) a 10 L cylindrical tube filled with sand (similar to the substrate that was used in the broodstock holding tank).Eggs fertilized on glass microscope slides were spread out onto several slides to limit egg density.All of the different fertilization substrates were then placed into five 50 L tanks with the same water as that used in the broodstock tank (Figure 4).Tanks were covered with blinds to avoid direct exposure to light.The light delivered by the overhead lamps was measured using an Ocean Insight Flame-S spectrometer.The total irradiance (400-700 nm) measured over the water surface of the tanks was on average 200 μW/cm 2 .Water flows through the tank was 10 L hour À1 during the embryonic period and was increased to 50 L hour À1 when the larvae reached the post-flexion stage.Water parameters were monitored daily to ensure stable rearing conditions.Oxygen was maintained at >80% of saturation, and the temperature was 7.8 C ± 0.5 C. The fertilization rate was calculated at 48 hours post-fertilization (hpf) by visual inspection of the different substrates under a microscope.Only a visual estimation of the fertilization rate was given for the eggs placed on the 500 μm filter mesh because moving or agitating this egg incubation substrate increased the chance of detaching eggs from the substrate and losing them in the rearing tank.
The efficacy of the different fertilization substrates was assessed using the following criteria: efficiency and ease of using the substrate, ease of observing and sampling embryos, bacterial growth, hatching success, and survival of hatched larvae.using peristaltic pumps to keep the concentration of prey above 3000 nauplii/L.

| Developmental timing and morphology of embryos and larvae
Embryogenesis was followed from the zygote stage to the free-swimming stage by examining embryos and larvae (n = 20-30 individuals per spawning from 8 females) under an Olympus SZX10 stereomicroscope coupled to a Moticam 1080 digital camera (Motic, Richmond, BC, Canada).The stereomicroscope was equipped with a thermally regulated microscope stage (Brook Industries, Lake Villa, IL, USA) set at 8 C, ensuring a constant temperature during observations.Major developmental landmarks and the morphology of specimens were observed.A small number of embryos were manually dechorionated to provide clear observations of the embryonic structures.Images and 20s videos were digitized using Motic Live Imaging Module.Heart rate (beat min À1 ) was determined by manually counting the number of ventricular contractions in each 20s video collected throughout development (n = 5 embryos per stage).ImageJ software 41 was used to measure chorion diameter (mm), volume of lipid droplet (mm 3 ), and standard length (mm; SL).
All of the developmental landmarks and morphology were gathered from data on sandeel collected during 2021 and 2022 and are reported in degree days (DD) and in days post-fertilization/hatching.Degree days were calculated by adding the mean daily water temperature in degrees Celsius for the total number of days measured.The developmental stages and both metrics are presented in Table 2.

| Histology
Sandeel embryos and larvae were sampled and fixed in 4% PBS buffered paraformaldehyde and stored at 4.0 C until further use.A histological evaluation for morphological structures was performed on 20-30 embryos at 60, 76, and 127 DD.The samples were dehydrated and embedded in paraffin using Histokinette Leica TP 1020 (Leica Microsystems Inc., Buffalo Grove, IL, USA).Thick serial sections (3 μm) were performed through the entire embryo using a Leica RM 2255 rotary microtome (Leica Microsystems Inc., Buffalo Grove, IL, USA).Histological sections were dewaxed and stained with hematoxylinerythrosin-safran (HES) following conventional histological procedures.Imaging was performed using a Leica DMRBE trinocular microscope (Leica Microsystems) connected to a SPOT insight CMOS camera (SPOT imaging, Sterling Heights, MI, USA).ORCID Prescilla Perrichon https://orcid.org/0000-0002-1410-878X

F
I G U R E 1 (A) Female lesser sandeel photographed after being stripped of eggs.The black arrow indicates the volume taken by the whole egg batch.(B) Egg batch placed on glass slide showing variation in developmental timing.The red arrow indicates grouped embryos with slower development (non-pigmented eyes) compared to a batch with a lower number of grouped embryos (pigmented eyes, black arrow).Photo credit: E. Sørhus.
0 and 3A).Brain differentiation is also well distinguished (Figures 2E 0 and 3A).At this stage, the tissue is characterized by immaturity indicated by a high level of blue hematoxylin staining (Figure3A).The cellular, tissue, and organ development accelerate at this point, and the sensory reflex begins with irregular peristaltic movements of the heart.The first muscle contractions of sandeel are observed at this point in development.At 76 DD, the finfold is present in its primary form and grows further with development.(Figure2F-F0 ).Heart chambers are discernible in videos with contracting atrium and ventricle (HR = 48 ± 1 beat.min À1 ; Figures 2F 0 and 3B).Embryonic movement increases sharply when they are touched from 76 DD.Heart rate frequency increases with further development reaching 72 ± 2 beat.minÀ1 at 174 DD.

F I G U R E 3
Light microscopic micrograph of lesser sandeel (Ammodytes marinus) embryo.(A) 60 DD/7 dpf; the tissue is characterized by immaturity indicated by high level of blue hematoxylin staining.Early eyes and brain were present.(B) 76 DD/9 dpf; the heart is present, and the somites contain both nuclei and striated muscle fibers.(C) 127 DD/16 dpf; rigid tissue structures made the sectioning challenging and result in a large degree of tearing tissue.Eyes structure with lens and iris are present.Long muscle fibers are visible along the trunk.b, brain; E, eye; h, heart; L, lens; M, muscle; ms, medulla spinalis; pcl, photoreceptor cell layer; r, retina; s, somite; uc, unvacuolated chordocytes, vc: vacuolated chordocytes, y: yolk.

From 1 -
day post-hatch (dph), larvae were fed a mix of Balanus crenatus nauplii (CryoPlankton, Planktonic AS, Trondheim, Norway) and Acartia sp.nauplii (hatched from resting egg, CFEED AS, Trondheim, Norway).Live algae (Rhodomonas sp.) were added as a food supplement until 30 dph.The larvae were then fed with live Artemia nauplii (INVE aquaculture inc., Salt Lake City, UT, USA) and frozen Artemia nauplii from 130 dph.Live feed and algae were pumped into the culture tanks continuously F I G U R E 4 Fertilization substrates placed in 50 L tank for eggs incubation of lesser sandeel (Ammodytes marinus).(A) filter mesh with pore sizes of 100, 300, or 500 μm.(B) Meshed bottom covered with 5 cm sand.(C) Basket with glass microscope slides.