Contribution to the reproductive ecology of Notoscopelus resplendens (Richardson, 1845) (Myctophidae) in the Central-Eastern Atlantic

Notoscopelus resplendens is an abundant myctophid in the region of the Central-Eastern Atlantic. As with a majority of other myctophid species, this species performs vertical migration, playing a key role in the oceanic food web and in carbon sequestration. We examined the reproductive biology of N. resplendens based on 579 specimens caught between 1997 and 2002 off the Canary Islands. We found that the maximum standard length (SL) was lower than the size reported by other authors. The sex ratio was not different from 1:1. The average size at first maturity (L50) was higher in females (60.34 mm SL) than in males (56.61 mm SL). The gonadosomatic index (GSI) at 50% sexual maturity in females was higher than that in males. The reproductive activity was observed from January to April, while from May onwards, the majority of fish caught were in the process of maturation. The macroscopic scale of maturation was validated through the histological analysis of the ovarian development. The batch fecundity was related to the standard length, with an average of 1068.69 ± 369.84 eggs/spawn. These first data obtained for N. resplendens indicated that it is a batch spawner with asynchronous ovarian development.


Material and methods
The study was based on the analysis of 579 specimens of N. resplendens caught during 4 cruises of the B/E "La Bocaina" between 1997 and 2002 (because the cruises did not cover all months of all years and to facilitate data analysis, the samples were grouped by 4-month periods, assuming no significant variation among years), off the Canary Islands (Central-Eastern Atlantic) (Fig. 1). The fishes were caught with a commercial semi-pelagic trawl net with a cod-end with 5 mm mesh size, but on the last cruise (2002), this mesh size was increased to 10.4 mm 49 . The hauls were conducted horizontally during the diurnal and nocturnal periods at a depth range between 13 and 1577 m. Fishing operations were monitored using acoustic telemetry, with a net-sounder SCANMAR, which provided information on the depth and the vertical and horizontal opening of the trawl mouth. Characteristics of the vessel and the net and a description of the fishing operations are given in other published works 50,51 .
Captured fishes were identified to the lowest possible taxon and stored in 70% ethanol for later analysis. As proposed in a previous study 44 , the number of gill rakers was used to identify individuals of the Notoscopelus genus. Once in the laboratory, for each fish, the total length (TL, 0.01 mm) and standard length (SL, 0.01 mm) were recorded using a digital calliper, and the whole body weight (BW, 0.01 g) and gonad weight (GW, 0.0001 g) using a digital balance. Generally, this species is captured using a midwater trawl net with a reduced mesh size, causing the loss or breakage of parts of the organisms' bodies (such as the caudal fin). For this reason, a consensus has been reached to use the standard length measurement (SL, from the tip of its nose to end of its last vertebrae, i.e., excepting the caudal fin). Here, the SL-TL relationship was estimated for all organisms using a linear equation to transform the published data of other authors (expressed in TL) and can be compared with our results. Individuals were then dissected, and the digestive tract, stomach and gonads were removed.
The specimens were grouped into length classes of 10 mm increments, and the length frequency distribution was analysed as percentages by 4-month periods [January to April (n = 191), May to August (n = 233) and September to December (n = 155)].
The relative fatness (energy reserves) and its relation to reproduction were estimated using the condition factor (K) 52 , which, as a measure of the physiological changes that occur in the organisms, was calculated for each individual as K = (BW/SL 3 ) × 100.
Sex was estimated from 518 samples due to external body dimorphism. Males present a large supra-caudal gland that allows them to be differentiated externally from females. The sex was confirmed after dissecting the fish and observing the gonads macroscopically. However, the state of maturity was determined for only 459 individuals through gonad macroscopic observation, following the classification criteria for fishes 53 to classify these stages as immature, developing, spawning capable, regressing and regenerating. The macroscopic scale of maturity was validated with histological analysis, including all maturity stages, considering the standard terminology for describing reproductive development in fishes 53  www.nature.com/scientificreports/ applied. For each macroscopic ovary stage, 10 gonads each were selected and processed histologically for each ovary maturity development stage. The sex-ratio was calculated for the whole sampling period and for the three 4-month periods (quarters) considered, and whether these rates were significantly different from the theoretical ratio of 1:1 was estimated using a chi-square test (χ 2 -test). For mature individuals (n = 187), the maturity ogives, lengths at first (L 50 ) and massive maturity (L 95 ) for both sexes, and the percentage accumulated by length class of mature individuals were calculated. The data obtained were fitted to a normal cumulative curve by iterative nonlinear regression. Furthermore, a comparison of means test was performed to detect possible differences between the L 50 of males and females. The data were fitted to a sigmoid function as follows: where Pr is the percentage of sexually mature individuals, r is a constant that indicates the slope of the curve, L 50 is the length at which 50% of individuals were mature, and L is the fish length for which the Pr is calculated.
To determine the spawning season, the gonads of individuals were monitored over time, and the average values of the Gonadosomatic Index (GSI) were calculated for each specimen, as the relationship between GW and BW-GW 54 : Another method used to determine the spawning season is based on the temporal variation in the frequency of individuals in each stage of maturity, which was determined by macroscopic observation of the gonads. To do this, the number of fish in each stage of maturity and their frequencies were calculated. To estimate GSI at 50% maturity (GSI 50 ), a logistic function was fitted to the fraction of mature fish per 0.5 GSI interval, for males and females, using a nonlinear least-squares regression. The logistic equation was: where PM GSI is the percentage mature at GSI, a is the slope, and b is the GSI 50 . The minimum size at maturity of males and females was taken as the smallest specimen with a GSI over GSI 50 based on the GSI-SL relationships 40 .  www.nature.com/scientificreports/ A total of 50 ovaries were fixed and preserved in 4% buffered formaldehyde for histological analysis to verify the previously assigned macroscopic maturity stages. For this, the fixed tissues were dehydrated in a series of ethanol solutions, cleared in isoparaffin H, and then embedded in paraffin in a vacuum chamber. Slices of tissue were sectioned at 4 µm and stained with Harris haematoxylin followed by eosin counterstaining 55 . To describe the scale of gonadal maturity, the standardized nomenclature 53 was used. The size of the oocytes within the four ovarian developmental stages were determined to describe the type (synchronous or asynchronous) of maturation and spawning based on ovaries that were previously processed histologically. The oocytes were removed from the ovary then placed in a vial and pipetted vigorously or full separation. The contents of the vial were then poured into a counting chamber, and the isolated oocytes were measured. The first 100 oocyte diameters were measured to the nearest 0.01 mm and classified based on histological correspondence images.
Finally, 84 samples of female gonads were collected to estimate the batch fecundity (BF) 56,57 through the gravimetric method, which was calculated for each female as the number of oocytes per unit weight multiplied by the total ovarian weight 40 .The proportion of the subsample from which the oocytes were extracted was evaluated with a target coefficient of variation (CV) of oocytes per unit weight of less than 5% 58 . The oocytes were manually released from the ovarian stroma and then counted using a stereoscopic microscope. Ovaries that did not contain early stage postovulatory follicle complex (POFs) were used because the presence of these indicate that some eggs have been already released 59 . In these ovaries, the oocytes at the most advanced stages, primarily Vtg3 and hydrated oocytes (H), were counted to estimate the batch fecundity 56,57 .
The statistical analysis was performed using the R programming language (V.3.6.0) 60 . Batch fecundity data from other studies are also shown for comparison purpose 40,61 . We obtained the values from the data points of figures using GetData Graph Digitizer V.2.26. The sampling map was generated using the geographic information system QGIS Development Team (V.3.12.1) 62 . ethical approval. The sampling was approved by the "Viceconsejería de pesca del Gobierno de Canarias" and the samples were obtained through commercial fishing trawls. We worked with preserved fishes in the laboratory.
Sampling and field studies. All necessary permits for sampling and observational field studies have been obtained by the authors from the competent authorities and are mentioned in the acknowledgements, if applicable. The study is compliant with CBD and Nagoya protocols.

Results
Length frequency distributions and sex ratio. The length frequency distribution showed the presence of two groups of lengths during the sampling period. From May to December, individuals of relatively small size (smaller than 60 mm SL) predominated, while larger fish (larger than 65 mm SL) were more abundant from January to April (Fig. 2).
Considering the entire sampling period, the sex ratio of N. resplendens was not different than 1:1 (1:0.92 Length at maturation. Among all fish sexed (n = 518), 36.10% of them were mature. Although the caudal gland in males was observed beginning at 25.29 mm SL, the L 50 was estimated to be 56.61 mm SL (n = 96). However, females (n = 91) reached the L 50 at a greater length (60.34 mm SL) (ANOVA, F = 4.829; p-value < 0.05). L 95 values were estimated to be 65.12 mm and 68.83 mm SL for males and females, respectively (Fig. 3).
Condition factor (K) and spawning season. K values ranged between 0.539 and 1.627 (Fig. 4) in the length range between 25 and 85 mm SL. Most fish captured from January to April had K values greater than 1, particularly in the length group larger than 60 mm SL. However, from May to December, these relatively high values of K were observed particularly in fish smaller than 60 mm SL. An ANOVA comparing the differences in K among sampling periods showed significant differences (F 0.05,2,>200 , = 17.12, p < 0.001), and the Post-Hoc Tukey test showed differences between January to April-May to August (p < 0.001) and May to August-September to December (p < 0.001). However, there were not significant differences between males and females in all sampling periods (F 0.05,2,>200 , = 0.218, p = 0.641) and for each period (January to April: F 0.05,2,>200 , = 0.141, p = 0.708; May to August: F 0.05,2,>200 , = 0.900, p = 0.344; September to December: F 0.05,2,>200 , = 1.298, p = 0.257).
Fish at stages III (Spawning) and IV (Regressing) (Fig. 5) were recorded during the whole sampling period, but from January to April (n = 134) and from September to December (n = 43), a relatively high proportion of spawning individuals were recorded (spawning takes place in winter and spring).
The analysis of the inner structure of oocytes in N. resplendence showed that Vtg1 oocytes were characterized by the presence of small vacuoles, the phases Vtg2 and Vtg3 were defined by the presence of vacuoles in the central and perinuclear areas of the oocyte, and of small yolk granules at its margin (Fig. 8).
Presence of oocytes in different videogenic state in the spawning capable phase determined that N. resplendens is a batch spawner with asynchronous ovarian development.
fecundity. Batch fecundity (BF) was significantly related to standard length in mature females (n = 84), increasing linearly for standard lengths between 66.6 mm and 82.39 mm SL (ANOVA, F 2,83 = 143.4 p < 0.01), with an average of 1068.69 ± 369.84 eggs by spawn pulse (range = 1089.24-1248.05) (Fig. 9a). As expected, the slope of the regression line of batch fecundity and SL of N. resplendens mature females was similar to those of Diaphus www.nature.com/scientificreports/ garmani and Diaphus pteretum, although the standard length of the latter two species was less than 60 mm SL (Fig. 9b). Statistically, the slopes for other species of Diaphus gender were significantly steeper than that of N. resplendens (ANOVA, F 4. 218 , F = 264.1; p < 0.01).

Discussion
Information about the distribution and some age and growth parameters of Notoscopelus resplendens have been studied in the Central-Eastern Atlantic 14,46,49,63,64 ; however, this is the first time that data have been reported for lengths at first (L 50 ) and full (L 95 ) maturity, spawning season and fecundity of this lanternfish species in this region (Table 1).   31 . Although the length range of individuals depends on the fishing gear used, in this study, the length range of N. resplendens is within the limits previously described for mesopelagic fishes (18.81 and 135 mm of SL) 12,16 .
The proportion of males and females in the entire sample was a 1:1 ratio, but we did not find any significant differences in their proportionality according to season. In this regard, the greater number of females observed in January to April could help maximize the egg-producing biomass 34,65 , and the opposite tendency recorded during May to August, when males predominated, could enhance the likelihood of mating but at the expense of a decreased number of egg producers in the population 66 . These differences in the sex ratio during the year (or years) and even with depth intervals have been observed in several species of myctophids, such as Benthosema pterotum 67 . www.nature.com/scientificreports/ The ratio of males to females decreased with SL, and most individuals in the largest class were females, as observed in many other myctophids 37,39,40 . This variation in the sex ratio with length might be related to faster growth and/or a longer life span of females in relation to males 65 . This has been documented in several myctophid species belonging to the Notoscopelus genus, where females grow faster than males and reach a larger maximum size 46,68,69 .
Previous studies have found that males N. resplendens begin to develop the large supra-caudal gland at 37.5 mm SL (41 mm TL), reaching sexual maturity at 59.65 mm SL (66 mm TL) 70 . Our data indicate that in the Canary Islands region this gland begins to develop at a smaller length than that reported by Hulley 13 (25.29 mm SL, i.e., 27.14 mm TL), and consequently, the average length of first maturity was also lower (56.61 mm SL, i.e., 61.60 mm TL) than previously estimated by the abovementioned author. Sexual dimorphism in luminous organs is known in many myctophids 71,72 . These caudal organs are considered to be related to sexual recognition in  www.nature.com/scientificreports/ myctophids because they can produce volleys of very fast and high intensity flashes 72,73 . Studies on B. pterotum 61 suggest that this bioluminescent sexual signalling might be used to facilitate communication between sexes at night, and this is possibly related to more efficient mating. In contrast, females of N. resplendens were mature at a larger size than males, with a length at first maturity of 60.34 mm SL. In other myctophid species, males mature at smaller sizes than females 29,30,34,65 . This would contribute to the greater biomass of mature females than males and to maximizing the egg-producing biomass.
In other studies of myctophid species, for example, in the Diaphus and Benthosema genera, the range of size at maturation is between 24.5 and 120 mm SL 38,40,61 , depending on the growth rate of the species and the region in which it is found. This could be the reason we found variability in the size of first maturity in the same species.
Seasonality in the spawning of many myctophid species has been reported in different areas of the world's oceans, although year-round spawning has also been observed 30,61,74 . In Bermuda 75 , N. resplendens spawns from winter to spring, with a peak of intensity in winter and early spring. The abundance of post-larvae suggested that spawning began prior to January and reached a peak in late February-early March. In a similar way, most of the individuals of N. resplendens caught off the Canary Islands showed relatively high values of K and GSI from January to April (corresponding mainly to winter and early spring), predominantly in those larger than 60 mm SL, but at the end of spring and summer (from May to August), the proportion of individuals with K lower than 1 was significantly greater, particularly among smaller-sized fish. In this regard, the K variations in N. elongatus kroeyeri were related to the spawning season 31 . Additionally, for the latter species, found in the Mediterranean 76 , the K values declined during the spawning season, which occurs from winter to spring, with a peak in April. Therefore, the lower K values observed in specimens of N. resplendens caught in March and May are related to the end of the reproductive season in the waters of the Canary Islands, during a similar period to that observed in the neighbouring waters of Mauritania, where spawning takes place in winter and spring 77 . Moreover, N. resplendens showed reproductive activity during winter and early spring (January to April), when the percentage of mature individuals was 70.2%, with reproductive activity occurring particularly in the larger length range.
In contrast, the histological analysis of the female gonads of N. resplendens showed oocytes in different developmental states, indicating that this species has asynchronous ovarian development with successive batch spawner seasons, as observed in other myctophids 28,38,39 . In addition, this species is characterized by intermittent spawning with long intervals between batches 77 . In the Canary Islands, batch fecundities vary from 578 to 2,122 eggs and progressively increase with fish length, from 66.6 to 82.39 mm SL. This correlation with SL has also been reported in other myctophid species 23,28,30,40,61 . However, although N. resplendens shows a standard length range greater than that in other myctophids, it has similar egg production, and consequently may have a higher energy cost for reproduction 39 . Although this study provides information on the reproductive biology of N. resplendens, we suggest that future research should study the fecundity and spawning season to understand the reproductive strategy of this species.
In the Canary Islands region, there is high fishing effort (artisanal and professional); however, although the fishes of the mesopelagic zones have no commercial interest, it is important to improve knowledge about this key fish group because they are prey for pelagic fishes (such as tuna) targeted by the fishing industry all along the West African coast. Because the myctophid species (such as N. resplendens) have a great abundance and biomass and because they are the link between secondary producers and upper trophic levels in the open ocean through the organismal migration 78 , they contribute significantly to the oceanic biological pump. In conclusion, the biological information provided for N. resplendens in this study (i.e., length frequency distribution, sex ratio, size at first maturity, spawning season and batch fecundity) together with information previously published about their age and growth 46 , contribute to expand the knowledge and the baseline for effective future management of this group of fish species.

Data availability
The datasets generated during the current study are not publicly available due to the data will be used by a student for a new degree research but are available from the corresponding author on reasonable request. Partial data was published in ''Catálogo de especies meso y batipelágicas. Peces, moluscos y crustáceos. Colectadas con arrastre en las Islas Canarias durante las campañas realizadas a bordo del B/E ''La Bocaina'' .