Nutritional vulnerability of early zoeal stages of the ornamental shrimp Lysmata ankeri (Decapoda: Caridea)

Abstract

The evaluation of the effects of early starvation and feeding on survival and growth in the early stages of the life cycle of ornamental marine caridean shrimp species is fundamental to establish adequate feeding protocols in their culture. In this study, we determine the nutritional vulnerability in the early larval stages of ornamental shrimp Lysmata ankeri exposed to different periods of starvation or feeding. The larvae were separated into three groups (zoea I-ZI, zoea II with ZI fed, and zoea II with ZI unfed) and subjected to two experiments: (1) point-of-no-return (PNR), comprising one or two days of initial starvation followed by feeding; and (2) point-of-reserve-saturation (PRS), comprising one or two days of initial feeding followed by starvation. Each experiment was still composed of two control groups: continuous feeding and continuous starvation. Larvae tolerated some periods of starvation, with a high PNR value (2.00) and low PRS (0.50). Longer periods of starvation influenced both growth and survival rates in zoea II stages. The nutritional vulnerability index for zoea I was 0.25, which represents a low dependence on food supply. In this study, it was observed that ornamental shrimp L. ankeri larvae hatch with energy reserves, presenting facultative primary lecithotrophy, in which they are able to moult from zoea I to zoea II using such reserves in the absence of food. In this sense, the early larvae stages (zoeas I and II) can tolerate a certain period of starvation, indicating the great potential of this species for aquaculture.

Introduction

In the last decades, the marine aquarium trade increased significantly and is considered a global multimillion dollars industry (Wabnitz et al., 2003; Murray et al., 2012; Rhyne et al., 2012, 2017; Leal et al., 2015; Chen et al., 2019) and with that, numerous species of marine fishes and invertebrates are the main focus of this trade (Lukhaup, 2002; Werner, 2003). Most marine aquarium species are still harvested from the wild, directly from coral reefs and associated habitats (Friedlander, 2001; Wood, 2001; Wabnitz et al., 2003; Smith et al., 2008; Rhyne et al., 2017). However, a sustainable alternative to avoid the capture of ornamental species would be the culture of the target species (Palmtag, 2017; Chen et al., 2019). Success in larviculture is particularly challenging since the early feeding of early larval stages appears to be the main bottleneck in commercial production of the culture of ornamental species (Danilowicz and Brown, 1992; Dhert et al., 2001; Van Eynde et al., 2019; Chen and Zeng, 2021; Groover et al., 2021). In this sense, some characteristics of larval development make it difficult to produce on a commercial scale, such as an extended larval phase, with many larval stages, resulting in a prolonged time to reach the juvenile stage and high mortality in culture systems of fishes and crustaceans (Calado et al., 2003a, 2003b). However, laboratory experiments indicate that nutritional stress during initial larval stages in ornamental species can cause high mortality rates and delay metamorphosis (Simões et al., 2002; Calado et al., 2005a, 2005b).

Among the invertebrates commercialized, one the most representative groups are crustaceans (Gasparini et al., 2005), in which caridean shrimp occupy, approximately, 40% of the trade in ornamental decapods (e.g., genus Alpheus, Cinetorhynchus, Gnathophyllum, Hymenocera, Lysmata, Rhynchocinetes, Thor) (Calado et al., 2008a; Rhyne et al., 2017). Lysmata species are among the most popular marine ornamentals and are of high retail value (Moe, 2001; Calado et al., 2003b; Dickson et al., 2020; Guéron et al., 2023). According to Rhyne et al. (2017), Lysmata ankeri Rhyne and Lin, 2006 is on the list top 20 live aquarium invertebrate species imported into the US by year, which represents about 2% of import volume (individuals as % total). Lysmata ankeri is found mainly in the United States, Florida, the Caribbean Sea, Haiti, Venezuela, Panama, Suriname, French Guiana, and Brazil (Rhyne and Lin, 2006; Alves et al., 2015; Barros-Alves et al., 2016). This species inhabits in coral reefs and rocky shores and are usually found at depths ranging from 5 to 15 m (Rhyne and Lin, 2006; Alves et al., 2015; Barros-Alves et al., 2016). According to Almeida et al. (2023) a longer larval development is expected for this species, with about nine to 11 larval stages. However, with regard to culture for this species, little information is available in the literature, such as reproductive cycle and embryonic development (Costa et al., 2021) and the importance of light and larval morphology in the starvation of newly hatched shrimps (Calado et al., 2008b). Other information is needed to help in the development of technologies for the culture of these species in the laboratory because most of the time, these ornamental crustaceans are obtained directly from the wild, in an unsustainable way, for the ornamental trade.

Thus, the evaluation of the effects of starvation on survival and size of the larvae in the early stages of ornamental species, such as L. ankeri, is important to establish adequate protocols of feeding in their culture. In some cases, newly hatched shrimp larvae have enough yolk reserves to allow them to moult from the first stage to the second in the absence of food, which presents facultative primary lecithotrophy (FPL) (see Anger, 1995, 2001; Thessalou-Legaki et al., 1999; Anger et al., 2015). At the other extreme, larvae commonly depend on exogenous food soon after hatching, being obligatory planktotrophic larvae (see Anger, 1995, 2001; Anger et al., 2015). In this context, Gebauer et al. (2010) proposed the Nutritional Vulnerability Index (NVI), which may be useful for examining the dependence of larvae on exogenous food. This index is estimated from values of ‘point-of-reserve-saturation’ (PRS) and ‘point-of-no-return’ (PNR). PRS represents the minimum time of initial feeding, after which enough energy reserves have been accumulated for the successful completion of the larval stage, independent of the later presence or absence of food, while PNR represents the time in which the individual loses the ability to change to the next larval stage after the nutritional stress caused by starvation in the initial period, even followed by feeding later (Anger and Dawirs, 1981; Paschke et al., 2004; Gebauer et al., 2010; Garcia et al., 2020). Despite the commercial importance of L. ankeri, there are still no studies that evaluate the nutritional vulnerability of this species. This information would assist in the design of protocols of culture, directly helping to reduce the effort spent capturing natural stocks. However, the aim of the present study was to determine the nutritional vulnerability in early larval stages of ornamental shrimp L. ankeri, exposed to different periods of starvation or feeding. Considering that the first zoeal stage of L. ankeri hatches with some yolk droplets (Costa et al., 2021), here, we expect that (1) larvae would tolerate periods of starvation and retain the ability to moult to the next stage (i.e., high PNR value and low PRS value); (2) higher survival rates; (3) higher carapace length (in mm) and shorter development time (in days); (4) there would be a low NVI value (i.e., low dependence on exogenous food).

Material and methods

Sampling of shrimps

Ten shrimps of L. ankeri were caught by collectors of ornamental organism from rocky bottoms in the Itapuã beach (12°57′26′′S, 38°21′14′′W), Salvador, northeast of Brazil, in June 2019, during night sampling, with a hand-net, among rocks, from 3 to 10 m deep. Collected specimens were individually placed in oxygenated plastic bags containing 200 ml of seawater from the collection site, in order to ensure survival during transport in thermal boxes to the Laboratório de Ecologia de Ecossistemas Aquáticos, Universidade Federal de Uberlândia, Minas Gerais, Brazil.

Broodstock maintenance

Ten specimens of L. ankeri in simultaneous hermaphrodite sexual phase were used to form five randomly assembled breeding pairs. The average carapace length (CL) of the specimens used was of 10.1 ± 0.8 mmCL. Each breeding pair was maintained in a rectangular aquarium (450 × 200 × 300 mm; 27-l capacity) containing plastic refuges. The aquariums were interconnected in a water-recirculation system adapted from Calado et al. (2007a) and Gregati et al. (2010). The water-recirculation system was equipped with an additional container for chemical and biological filtering, containing a protein skimmer (Marine Sources Red Devil Professional Skimmer – RDC 250), biofilters, and activated carbon (500 g). The shrimps were kept under a 12-h light-dark photoperiod. Artificial seawater was prepared using freshwater purified by a deionizer system and mixed with the salt Crystal Sea^® produced by Marine Enterprises International^® (Baltimore, MD, USA), following the instructions of the manufacturer. The mean temperature and salinity (± SD) were maintained at 25 ± 1°C and 30 ± 1 ppt, which correspond to the approximate values recorded in the natural environment. Temperature was controlled with a digital thermostat (±0.1°C accuracy), two heaters (H-606e, 300 W, HOPAR), and a chiller (1/3 hp³). The shrimp were fed ad libitum once a day in the morning with fresh fragments of shrimp and industrialized food (Thera^® New Life Spectrum, containing the primary ingredients, according to the manufacture: 38% of protein, 7% of fat, 5% of fibre, 10% of moisture, and 8% of ash).

Rearing of larvae and experimental conditions

Larvae of L. ankeri hatched from 7 to 10 days after egg laying, with newly hatched larvae (zoea I, ZI) being attracted to an LED light trap installed in the aquarium (for more details, see Gregati et al., 2010). In the morning, the aquarium was checked for newly hatched larvae, which were captured with a Pasteur pipette and individualized for the experiments.

Only the most active newly hatched larvae, which showed a positive phototactic response were used in the experiments. This criterion is adopted in previous studies on other decapod larvae (Calado et al., 2008b, 2010; Gregati et al., 2010). The above procedure was repeated until a total of 20 replicates (larvae) for each treatment, every time other female larvae hatched, to ensure the same larvae number for each female. Each larva was considered a replicate, following criteria from previous studies on other decapod larvae (Gebauer et al., 2010; Guerao et al., 2012; Pantaleão et al., 2015; Espinoza et al., 2016; Barros-Alves et al., 2018). Based on a pilot study, each experiment lasted three days, i.e., the time required for newly hatched larvae to reach the subsequent stage.

Larvae were reared individually in plastic bottles (80 ml) containing 50 ml of artificial seawater. The water used in the experiment was prepared using freshwater purified by a deionizer system and synthetic sea salt suitable for marine aquariums (Red Sea^®). The plastic containers were arranged inside a plastic tray (530 × 400 × 120 mm) filled with freshwater and maintained at 25 ± 1°C, using heaters (AT 180, 100 W, ATMAN). Daily, the plastic bottles were cleaned and a water from each larval container was also approximately 90% renewed by 1 μm filtered artificial seawater. Temperature and salinity were checked daily. In the experiments with feeding, newly hatched Artemia nauplii were provided as food at a density of 2 prey ml⁻¹, and any remaining food was removed the next morning. The larvae were checked daily in the morning to register death or moulting (evidenced by the presence of exuviae). At the end of the experiment, all larvae were conserved in ethyl alcohol 70% and carapace length (CL) was measured under a stereomicroscope (Leica M205 C) equipped with the imaging system Leica Application Suite (LAS) version 4.4 (accuracy of 0.05 mm). The CL was determined from the post-orbital angle to the posterior margin of the cephalothorax. Each treatment was concluded when the larvae reached the next, died, or did not moult to any stage by the last day of the experiment.

PNR and PRS experiments

A study revealed that newly hatched Lysmata species does not cannibalize each other, assuring that larvae which were not being provided larval preys were truly in continued starvation (Calado et al., 2007b). In these experiments, we tested the starvation resistance on survival, moulting, and the carapace length in the three groups of different individuals: (1) newly hatched Lysmata larvae (ZI); (2) second zoeal stage larvae produced with ZI fed (ZII with ZI fed), and (3) second zoeal stage larvae produced with ZI unfed (ZII with ZI unfed) (see Figure 1A).

Figure 1. Protocol used during food restriction experiments in the initial larval stages (zoea I and II) of Lysmata vittata. (A) Groups of larvae used in the PNR and PRS experiments (grey bars). (B) Starvation treatments to determine the point of no return (PNR). Larvae were initially starved (S) for the specified number of days and then fed for the remaining days of the experiment. (C) Feeding treatments to determine the point of reserve saturation (PRS). Larvae were initially fed (F) for the specified number of days and then starved for the remaining days of the experiment. CF, continuously fed control; CS, continuously starved control. (Experimental design based on Paschke et al., 2004).

Newly hatched larvae from at least three parental shrimps of L. ankeri, were collected after hatching, and these larvae were isolated in a small plastic container. The PNR experiment comprised different periods of initial starvation and subsequent days of continuous feeding: larvae starved for one and two days after hatching (Figure 1B). On the other hand, the PRS experiment comprised different periods of initial feeding and subsequent days of starvation as follows: larvae feeded for one and two days after hatching (Figure 1C). For both experiments, two control treatments were also considered: continuous feeding (CF) and continuous starvation (CS) (Figure 1B and C). Twenty replicates were used for each treatment, for a total of 8 treatments × 20 larvae (replicates) × 3 groups of different individuals, totalling 480 larvae. Each experiment lasted a maximum three days.

Data analysis and statistics

‘Moulting’ was determined as the percentage of individuals from each treatment that moulted before to the next stage. ‘No moulting’ refers to the percentage of individuals in a certain treatment that did not die and never moulted. ‘Mortality’ was calculated by the percentage of individuals that died before the experiment ended, such as described by Calvo et al. (2012). The development time in days (mean ± SD) was calculated for all larvae per treatment. The PNR₅₀ and PRS₅₀ indices were calculated from the mortality data of each treatment. The PNR₅₀ and PRS₅₀ values were calculated by adjusting the sigmoidal Boltzmann model described in the following equation:

$$M = \displaystyle{{A_1-A_2} \over {1 + e^{( {x-x_o} ) /d_x}}} + A_2, \;$$

where M is the percentage of ‘mortality’ + ‘no moulting’ individuals, A ₁ is the minimum value of M, A ₂ is the maximum value of M, x ₀ is the time (days) for M to reach 50%, d _x is the time constant, and x is the time (days) of initial starvation (PNR) or feeding (PRS) (Paschke et al., 2004; Bas et al., 2008; Gebauer et al., 2010). The values of PNR₅₀ and PRS₅₀ were estimated considering the development time of ZI to ZII, of ZII produced with ZI fed to ZIII and of ZII produced with ZI unfed to ZIII. The PRS₅₀/PNR₅₀ quotient is calculated to determine the NVI, in which values range from zero to infinity. NVI values lower than 0.5 indicate that larvae present low dependence on food exogenous; while NVI values higher than 1.0 indicate that larvae are dependent on exogenous food (see Gebauer et al., 2010).

Parametric tests were used when data met the model assumptions, otherwise, equivalent non-parametric tests were used. The model assumptions of homoscedasticity (Levene's test) and normality (Shapiro–Wilk's test) were initially tested (Zar, 2010). The chi-square (χ²) test was used to test differences in larval ‘mortality’ + ‘no moulting’ between the treatments and the CF control. A one-way ANOVA (parametric) test was used to test differences in the development time, followed by a Tukey's test for multiple comparisons among treatments and the CF control. A linear regression was applied to analyse differences in the carapace length between CF control and other treatments. Linear regression equations were obtained after log-normal transformation of the data. The level of significance was set at 5% for two-tailed P-values (Zar, 2010). All statistical analysis was performed using R Statistical (version 4.0.5; R Foundation for Statistical Computing, Vienna, Austria).

Results

PNR experiment

In the PNR experiment, larvae in ZI reached the ZII stage without significant difference in survival among treatments (Figure 2A). The time required to ZI moulted to ZII was 2.11 ± 0.32 (mean ± SD) days (Table 1) in the control treatment (CF) and also did not differ compared to any of the other treatments (ANOVA; F = 1.45; d. f. = 3; P = 0.238). However, the size of ZII was different depending on the initial starvation time to which larvae in ZI were exposed (Tukey; P < 0.001), with a negative relationship between initial starvation time and larvae size (Linear regression; F = 42.47; d. f. = 58; r ² = 0.428; P < 0.001) (Figure 2B and Table 1). For the ZII with ZI fed, survival to ZIII only differed in the CS control (χ², P = 0.041) (Figure 2C). A pattern similar to that recorded for the ZI was recorded for the ZII with ZI fed, in which no significant effect was observed on development time (ANOVA; F = 1.22; P = 0.312) (Table 1), but with a significant effect over size (Tukey; P < 0.001), in which CL decreased while starvation days increased (Linear regression; F = 58.42; d. f. = 55; r ² = 0.515; P < 0.001) (Figure 2D). Finally, for the ZII with ZI unfed ZI) group, the effects of starvation are significant, and survival until ZIII depends on continuous feeding during the ZII stage (Figure 2E). All larvae in this group, exposed to some starvation (S1, S2, and CS treatments) did not reach ZIII or died mostly, therefore, the effect of starvation on size could not even be tested (Figure 2F).

Figure 2. Point of no return (PNR) of Lysmata vittata. (A) Percentage of moulting, no moulting and mortality of zoea I in different treatment groups; (B) Carapace length (in mm) of zoea I under different days of starvation; (C) Percentage of moulting, no moulting and mortality of zoea II with ZI fed in different treatment groups; (D) Carapace length (in mm) of the zoea II with ZI fed in the different days of starvation; (E) Percentage of moulting, no moulting and mortality of zoea II with ZI unfed in the different treatment groups; (F) Carapace length (in mm) of the zoea II with ZI unfed in the different days of starvation. The larvae were initially starved (S) for the specified number of days and then fed for the remaining days of the experiment (see text for details). CF, continuously fed control; CS, continuously starved control. *indicate significant differences compared to the CF control (χ², P < 0.05).

Table 1. Average development time (in days) (±standard deviation) and carapace length (in mm) of zoea I (ZI), zoea II with ZI fed or zoea II with ZI unfed of Lysmata ankeri in different point of no return (PNR) treatment groups

CF, continuously fed control; CS, continuously starved control.

Larvae were initially starved (S) for the specified number of days and then fed for the remaining days of the experiment (see text for details).

Different letters indicate significant differences among treatments (Tukey, P < 0.05) (n = 20 replicates of larvae per treatment, per each stage of zoea).

PRS experiment

In the PRS experiment, similarly to the PNR experiment, the group of larvae in ZI reached ZII with mortality that did not differ between treatments (Figure 3A). The time for ZI to moult to ZII also did not differ significantly between treatments (ANOVA; F = 0.11; d. f. = 3; P = 0.954) (Table 2), but the size of ZII was affected by the starvation time in ZI (Tukey; P < 0.001) (Table 2), here, with a positive relationship between size and the number of days on which the food was initially available (Linear regression; F = 58.35; d. f. = 65; r ² = 0.473; P < 0.001) (Figure 3B). For the ZII with ZI fed, only of the CS treatment did not reach ZIII in the same proportion as those from the CF treatment (Figure 3C). The development time of larvae from ZII with ZI fed to ZIII also did not differ significantly between treatments (ANOVA; F = 0.41; d. f. = 3; P = 0.745) (Table 2). However, a significant effect of starvation on the size of ZIII was recorded, with CS larvae being smaller in ZIII compared to larvae exposed to all other treatments (Table 2). The positive relationship between size and the number of days in which food was initially available was also recorded for this group of larvae (Linear regression; F = 58.42; d. f. = 55; r ² = 0.515; P < 0.001) (Figure 3D). Finally, 100% of larvae from the CS treatment, from the ZII with ZI unfed did not reach ZIII. However, if the food is provided at least one day, the survival until ZIII does not differ significantly from CF treatment (Figure 3E). The development time of larvae until ZIII did not differ between treatments (ANOVA; F = 0.89; d. f. = 2; P = 0.420) (Table 2), nor did the size of larvae that reached ZIII differ between treatments (ANOVA; F = 0.52; d. f. = 2; P = 0.597) (Table 2). The same positive relationship between size and the number of days on which food was available initially observed for ZII with ZI fed was recorded here for ZII with ZI unfed (Linear regression; F = 14.64; d. f. = 38; r ² = 0.278; P < 0.001) (Figure 3F).

Figure 3. Point of reserve saturation (PRS) of Lysmata vittata. (A) Percentage of moulting, no moulting and mortality of zoea I in different treatment groups; (B) Carapace length (in mm) of zoea I under different days of feeding; (C) Percentage of moulting, no moulting and mortality of zoea II with ZI fed in different treatment groups; (D) Carapace length (in mm) of zoea II with ZI fed in the different days of feeding; (E) Percentage of moulting, no moulting and mortality of zoea II with ZI unfed in the different treatment groups; (F) Carapace length (in mm) of zoea II with ZI unfed in the different days of feeding. The larvae were initially fed (F) for the specified number of days and then starved for the remaining days of the experiment (see text for details). CF, continuously fed control; CS, continuously starved control. *indicate significant differences compared to the CF control (χ², P < 0.05). **indicate marginally significant differences compared to the CF control (χ², P = 0.059).

Table 2. Average development time (in days) (±standard deviation) and carapace length (in mm) of zoea I (ZI), zoea II with ZI fed or zoea II with ZI unfed of of Lysmata ankeri in the different point of reserve saturation (PRS) treatment groups

CF, continuously fed control; CS, continuously starved control.

Larvae were initially fed (F) for the specified number of days and then starved for the remaining days of the experiment (see text for details).

Different letters indicate significant differences among treatments (Tukey, P < 0.05) (n = 20 replicates of larvae per treatment, per each stage of zoea).

Nutritional vulnerability

The PNR₅₀ and PRS₅₀ mean (± SD) values of the early larval stages (ZI) obtained from the sigmoid curves were 2.00 ± 0.00 and 0.50 ± 0.00 days, respectively. The PNR₅₀ and PRS₅₀ mean (± SD) values of the ZII produced with ZI fed obtained from the sigmoid curves were 1.41 ± 0.00 and 1.00 ± 0.00 days, respectively. The PNR₅₀ and PRS₅₀ mean (± SD) values of the ZII produced with ZI unfed obtained from the sigmoid curves were 0.53 ± 0.00 and 0.58 ± 0.00 days, respectively. From these values, the NVI of ZI, ZII produced with ZI fed and ZII produced with ZI unfed was estimated to be 0.25, 0.71, and 1.09, respectively.

Discussion

The nutritional vulnerability in early larval stages of ornamental shrimp L. ankeri exposed to different periods of starvation or feeding was determined here. Thus, we verified that this species is able to develop through the early larval stages in the absence of exogenous food. The results support our expectations that (1) larvae would tolerate periods of starvation and retain the ability to moult to the next stage (i.e., high PNR value and low PRS value); (2) higher survival rates; (3) higher carapace length (in mm) and shorter development time (in days); (4) there would be a low NVI value (i.e., NVI ≤ 0.5).

Larvae of L. ankeri submitted to PNR treatments moulted from zoea I to II and, when fed on zoea I, they moulted to zoea III. In the CF treatment, the mortality rate and development time do not differ from the other treatments, except for ZII produced with ZI unfed in CS treatment. In the study of the Calado et al. (2007b), for some species of Lysmata, 100% of the larvae reached the subsequent stages when fed continuously. In this study, the mortality rate did not differ between treatments, but 100% survival was not observed in CF control, which may be attributed to the effect of starvation in the first hours after hatching, considering that the larvae were captured only in the morning. On the other hand, larvae submitted to PNR treatments, a negative influence on larval size and development time of L. ankeri was observed, i.e., larvae submitted to PNR treatments were significantly smaller. In cases where the initial starvation period increased, the efficiency of the feeding activity might be affected, and reserves might not be absorbed or accumulated, resulting in smaller larvae, even that they are re-fed for a longer period (Liddy et al., 2003; Stumpf et al., 2010). According to Knowlton (1974) and Giménez and Torres (2002), larvae prioritize their energy for growth, i.e., situations where suboptimal or stress conditions, the animal survives extending the time of the larval stage but does not grow. Moreover, our results indicated that the damage caused by starvation during the early larval development might be irreversible. This suggestion is supported in the study by Barros-Alves et al. (2020), in which larvae that were not fed during the first day died or did not moult to the next stage. Periods of starvation also negatively affected the body size of the Neocaridina davidi (Bouvier, 1904) (Pantaleão et al., 2015) and Lysmata vittata (Stimpson, 1860) (Barros-Alves et al., 2020), in which larvae smaller than those exposed to continuous feeding were observed for the newly hatched stages, when exposed to starvation immediately after hatching. Other studies have also shown that the size of the larvae may be smaller when subjected to the starvation period and these studies point to a direct relation between starvation time and the irreversible effects in the larvae (e.g., larval size) (Mikami et al., 1993; Giménez and Anger, 2005; Calado et al., 2005a; Figueiredo et al., 2008).

Larvae of L. ankeri submitted to PRS treatments moulted to the next stage of development in all cases, except for ZII produced with ZI unfed in CS treatment. Larvae fed only in the early days (e.g., F1 and F2) and, subsequently, exposed to periods of starvation moulted to stage III with a similar time of development and mortality rate when compared with larvae fed continuously. Similar results were observed for L. boggessi and L. seticaudata which also reached the next stage, when submitted to PRS treatments (Calado et al., 2007b). According to Romero-Carvajal et al. (2018), L. boggessi larvae hatch with ~10–20% of the yolk distributed in the cephalothorax, that may justify the independence of exogenous food during the early larval stages. The same occurs with L. vittata larvae, which hatch without macroscopically visible yolk reserves, indicating that the yolk must be mostly consumed during the embryonic development (Alves et al., 2019), and L. ankeri, which larvae hatch only with small visible droplets of yolk (Costa et al., 2021). Thus, this result also corroborates with larvae of L. boggessi Rhyne and Lin, 2006 by Calado et al. (2007b) and L. seticaudata (Risso, 1816) by Calado et al. (2008b), which were not fed in the first days of the development. According to Calado et al. (2007b), larval survival may be related to the feeding of the breeding female during the period of gonad maturation, that may affect embryo development, and this factor cannot be controlled in the laboratory.

In this study, we observe that the ornamental shrimp L. ankeri presents low dependence on exogenous food in the early larval stages (NVI = 0.25). According to Gebauer et al. (2010), NVI ≤ 0.5 represents low dependence on food supply, i.e., low nutritional vulnerability. On the other hand, feeding independence was observed for other larvae of decapod crustaceans, in which PRS were lower than PNR values (NVI < 1), such as Neohelice granulata (Dana, 1851) studied by Giménez (2002) and Bas et al. (2008); Crangon crangon (Linnaeus, 1758) by Paschke et al. (2004); Maja brachydactyla Balss, 1922 by Guerao et al. (2012), N. davidi by Pantaleão et al. (2015), Ucides cordatus (Linnaeus, 1763) by Souza et al. (2018), and L. vittata Barros-Alves et al. (2020). However, many decapod larvae present PRS values greater than PNR (NVI > 1), and are completely dependent on exogenous food, such as: Armases cinereum (Bosc, 1801) studied by Staton and Sulkin (1991); Petrolisthes laevigatus (Guérin, 1835) by Gebauer et al. (2010); Grimothea monodon (H. Milne-Edwards, 1837) by Espinoza et al. (2016); Panulirus argus (Latreille, 1804) by Espinosa-Magaña et al. (2017) and Stenorhynchus seticornis (Herbst, 1788) by Barros-Alves et al. (2018). It is worth referring that exogenous food dependence may be variable among larval stages, since L. ankeri showed an intermediate level of food dependence in ZII produced with ZI fed (NVI = 0.71). This result was corroborated in the study by Anger and Dawirs (1981), in which Hyas araneus (Linnaeus, 1758) showed food independence during the first larval stage (NVI < 0.5), and later (zoea II) presents high dependence (NVI > 1.0).

In this study was observed L. ankeri larvae hatching with high energetic reserves, displaying FPL (i.e., being able to moult from zoea I to zoea II using internal energetic reserves in the absence of food). In this sense, larvae can tolerate a period of starvation, once they had access to exogenous food after hatching. FSL was also recorded for L. ankeri in the study by Calado et al. (2008b). FSL is commonly observed in other shrimp of the genus Lysmata with ornamental interest, such as L. boggessi studied by Calado et al. (2007b), L. seticaudata by Calado et al. (2007b) and Calado et al. (2008b), and L. vittata by Barros-Alves et al. (2020). In this case, according to Anger (2001) and Espinoza et al. (2016), the larvae can hatch with parts of the egg yolk deposited in the form of lipid droplets reflected by increasing fat in the hepatopancreas, which are responsible for nutrition in this initial phase.

On the other hand, larvae of L. amboinensis (De Man, 1888) and L. debelius Bruce, 1983 showed a lower tolerance to starvation periods (Calado et al., 2007b), since these larvae fully depend on exogenous food to complete their development and they are fully planktotrophic. The dependence on exogenous food by newly hatched larvae has been also recorded for other decapods, such as M. brachydactyla by Guerao et al. (2012), C. crangon by Paschke et al. (2004) and S. seticornis by Barros-Alves et al. (2018). For these planktotrophic species, feeding during the first larval stage is crucial for larval development, since irreversible damage can be caused in the absence of food during this larval stage.

For aquaculture purposes, it is important to highlight than despite L. ankeri featuring FPL, live feeds should be presented to larvae immediately after hatching to avoid any unnecessary nutritional stress that may latter impact larval development. Although the survival rate and development time were not influenced, but the size of the larva was influenced by the lack of food, even with only one day of starvation. Thus, larval size (length and/or weight) is an indicator of growth, and the increased growth and reduced in size during larval stages were related to further growth to juvenile stages (Castille et al., 1993).

Conclusions

Overall, our results support that for the success in the culture of L. ankeri during the first larval stages, despite the existence of FSL, and having an energy reserve in the first larval stage, irreversible effects (e.g., in the size of the larva) for larviculture can be observed if larvae are not properly fed at all times. By supplying food immediately after hatching to larvae one safeguards: (1) higher survival; (2) shorter development time in days to reach the next stage; and (3) larger larvae being produced. All these factors together can contribute to success in the larviculture of this ornamental shrimp, although further studies are needed in order to provide key information on the larviculture requirements of this species.