Reproductive aspects of prawn Desmocaris trispinosa (Aurivillius, 1898) from a periurban stream of Abidjan (Banco stream, Côte d’Ivoire)

: Aim: Characterize habitats and determine the reproductive parameters of the prawn Desmocaris trispinosa (Aurivillius, 1898) from Banco stream, a periurban stream of Abidjan (Côte d’Ivoire, West Africa). Methods: Sampling was conducted monthly, from July 2020 to June 2021 at four stations. Environmental characteristics were measured before prawn sampling. The sampling was made with a dipnet during five-minute active fishing conducted by a single operator at each station. Results: According to environmental variables, sampling stations were divided into three groups: almost undamaged (B4), slightly disturbed (B1) and disturbed (B2 and B3). In total, 467 specimens of D. trispinosa were captured in which 127 males, 275 females and 65 juveniles. Overall, the sex ratio (1:2.1) was in favour of females. Considering stations, the sex ratio was also in favour of females. But, only in the almost undamaged station (B4), the difference was significant ( χ 2 = 70.20; P<0.001). At this station, a variation of sex-ratio was observed in two seasons. In total, 116 ovigerous females were analysed on all stations. Fecundity varies from 5 to 16 oocytes with an average of 11.03 ± 2.18. First maturity size was estimated at 4.82 mm carapace length for the combined data. Females from the station free from human activities (B4) reached sexual maturity (4.23 mm) before those of the other stations (4.97 in B1 and 5.70 mm in B2). D. trispinosa from Banco stream reproduces continuously year-round with two peaks in the rainy season. The oocyte volume was negatively influenced by temperature and positively by dissolved oxygen. Conclusions: D. trispinosa reproduces year-round in undisturbed stations and reaches sexual maturity faster at these stations compared to disturbed stations. The sex ratio was in favour of females. Consequently, this species would be naturally polygamous.

1898) was the largest distributed and the most abundant in the stream.It was sampled both in undisturbed and disturbed habitats.This small species seems to tolerate unfavourable conditions such as low dissolved oxygen and significant organic matter (Cumberlidge, 2006a).To our knowledge, with the exception of the research of Powell (1979) on the larval development of D. trispinosa, there is a lack of scientific data on the reproductive biology of this small prawn.However, reproduction is one of the important aspects in both ecology and biology of species (Zare et al., 2011).In addition, knowledge of reproductive biology characteristics of species is an essential tool that can provide relevant information for their conservation and their preservation (Lévêque & Paugy, 2006).The present study aims to examine the reproductive aspects of the prawn D. trispinosa in Banco stream.The specific objectives of this study are (i) to characterize the habitats of D. trispinosa through physico-chemical parameters and (ii) to determine the reproduction parameters (sex ratio, eggs diameter and volume, size at first sexual maturity), including the breeding period in the Banco stream basin.Results from this study constitutes preliminaries data on reproductive biology characteristics D. trispinosa.

Introduction
Prawn are essential species in river food webs because they serve as intermediate consumers, linking periphyton and detritus production to higher trophic groups (Browder et al., 1994;Frédérick & Spalding, 1994).Together with crabs, they constitute the majority of the macroinvertebrate's biomass in hydrosystems (Boulton & Lloyd, 1991;Sheldon & Walker, 1998).They are also an important food resource for carnivorous fish (Resende et al., 1996).Therefore, they are often considered key animals, especially in small streams (Pringle et al., 1993).
Four stations were chosen on the stream; three stations (B1, B2 and B3) were located on the main channel, while station B4 was chosen on a tributary of the stream.Station B1, located upstream of the river, is characterized by relatively turbid water, and the presence of fragmented leaves, woody debris, riparian vegetation (Turraenthus africanus, Petersianthus macrocarpus, Dacryodes klaineana and Thaumatococcus daniellii), with a sandy substrate.It should also be noted that this station is partially disturbed probably due to the arrival of a small amount of runoff water causing silting of the bed.Station B2 is situated near the forestry school and is characterized by a turbid water, a muddy, sandy and clayey bottom with marginal grassy vegetation composed mainly of Cyclosorus striatus and Nephrolepis biserrata.This station is also the receptacle for wastewater from the forestry school and dwellings, and runoff water from Abobo city.As for station B3, it is located near the mouth of the river and is characterized by the presence of Indian bamboo trees debris y with a clay and gravelly bottom.The station B4 is chosen on a permanent tributary of the Banco stream.At this site, the water is crystal clear and is free of any anthropogenic pollution.The substrate is mainly composed of mud and vegetal debris.

Data collection
Sampling was carried out monthly from July 2020 to June 2021, except for October 2020.Measurements of all environmental variables were made in situ prior to shrimp sampling.Electric conductivity (µS.cm −1 ), water temperature ( • C), dissolved oxygen (mg.L −1 ) and pH were measured using a multiparameter (HANNAH Hi 9828).The turbidity (NTU) was measured using a turbidimeter (AQUALYTIC CD 24).The depth (m) and width (m) of the water were measured (average of five trials) to the nearest centimeter using a tape measure.
The sampling of D. trispinosa was identical at all the stations on the stream.Shrimp were dipnet-caught during a five-minute active fishery by a single operator at each site.Within the framework of our study, the dip-net used has a metallic opening of 50 cm in diameter with a depth of 30 cm; the opening is fixed to a sleeve two (2) m long which allows to hold the material.After capture, ovigerous females (females carrying eggs) were placed in separate plastic jars to prevent loss of eggs during transport.These jars were labelled and contained 70% ethanol.In the laboratory, sexual differentiation was made through the observation of the endopodite morphology of the second pair of pleopods, as proposed by Ismael & New (2000).Prawn individuals were identified according to Monod (1966) and Cumberlidge (2006b).All specimens clearly recognized as D. trispinosa were counted and then measured.The total body length (TL: measurement taken from the tip of the rostrum to the tip of the telson) and the carapace length (CL: measurement taken from the orbital hollow to the middle of the posterior end of the cephalothorax) were measured using a manual electron microscope (Celestrom 2.0).Shrimp weights were determined with a Baxtran electronic scale (model ANG200A1, capacity 200g) to the nearest 0.001g.Individuals with broken rostrum or another part of the body destroyed were discarded.

Determination of reproduction parameters
Size at first maturity was estimated by fitting the frequency of mature females to a logistic model (King, 2007): where r is the slope, TL is the total length and Lm the size at first maturity (  account the integrity of the eggs and the months of collection.The entire egg mass was removed from the abdomen, separated using forceps, and all eggs counted using an electron microscope (Celestrom 2.0).Egg development stages were classified according to the criteria proposed by Lara & Wehrtmann (2009): Stage I: Egg-bearing females with orange eggs, recently extruded egg, yellow uniform, no visible eye pigments; Stage II: Eggbearing females with brown eggs, barely visible eggeye pigments; Stage III: Egg-bearing females with grey eggs, clearly visible and fully developed eyes.
The sex ratio was obtained by identifying the number of males (M) and females (F) by station and by sampling season at each station using the formula: Oocyte diameters were measured in the middle (d1) and apical (d2) portions of the eggs using an electron microscope (Celestrom 2.0).Since the oocytes are ovoid, the diameter was calculated with the formula , where d1 represents the small diameter and d2 is the large diameter (Budi et al., 2020).The formula calculation of oocyte volume (V) is that proposed by Turner & Lawrence (1979) and used by several authors (e.g.: Cartaxana, 2003;Béguer et al., 2010): The reproductive period was determined through analysis of the relative frequency of ovigerous females monthly at each sampling site as implemented by Fransozo & Mantelatto (1998), and Hoffmann & Fransozo (2010).

Data analysis
A principal component analysis (PCA) was used for the ordination of the stations according to the physico-chemical characteristics.The Student t test was used for seasonal (dry and rainy) variation of oocyte diameter and oocyte volume.The one-way analysis of variance test (ANOVA I) was performed for spatial variation of fecundity (absolute and relative) and oocyte size (diameter and volume).Linear regressions were generated to describe the relationship between absolute fecundity and the size of ovigerous females; carapace length (CL) and absolute fecundity; fecundity and individual's weight.In addition, the linear regression was used describe the relationship between environmental variables and reproduction parameters.The chisquare test was used to compare the value of the sex ratio found with the theoretical sex ratio (1:1).The significance threshold for the probability value is p < 0.05.The R software was used to perform these two tests.

Abiotic differentiation of sampling stations
Principal component analysis (PCA) made it possible to establish the abiotic typology of sampling stations (Figure 2).The first two axes expressed 55.13% of the total variance: 33.5% for axis1 and 21.63% for axis2 (Figure 2A).The correlation circle (Figure 2B) revealed that axis1 was strongly and negatively correlated by width, conductivity and depth.On the other hand, axis2 is strongly and positively correlated by the pH and the turbidity and to a lesser degree the temperature.The factorial map (Figure 2C) distinguishes three groups of stations.The first group is composed of samples from stations B2 and B3.In these stations, water was generally characterized by the high values of electric conductivity, wet bed width and water depth.The second and third groups were composed of sample from single station (B1 and B4 respectively).The samples from B1 were distinguished by high values of pH, turbidity and water temperature.In contrast, samples from B4 were mostly differentiated by low values of these environmental variables.

Spatial and season variations of the sex-ratio
In the present study, 467 specimens of D. trispinosa were captured in the four sampling stations (B1, B2, B3 and B4); 127 males, 275 females and 65 juveniles, resulting in a sex-ratio of 1:2.1 (male:female) in favour of females (Table 1).The results of the adjustment test (chi 2 test) showed that the observed sex-ratio was significantly different from the theoretical ratio of 1:1 (χ 2 = 54.48,P<0.001).On all the stations, the Chi 2 values showed that there is no difference from the theoretical ratio of 1:1, except for station B4 whose sex-ratio was 1:4 in favour of females (χ 2 = 70.20;P<0.001).Seasonal variations in the sex-ratio were in favour of females over the entire study period, except for station B3 in the dry season (Table 2).However, only station B4 recorded sex-ratios that were significantly different from the theoretical sex ratio between the two seasons sampling (1: 6.1; χ 2 = 2.48, P<0.001 in the rainy season and 1:1.7; χ 2 = 0.85, P<0.05 in the dry season).

Fecundity
Overall, fecundity varied from 5 to 16 oocytes with an average of 11.03 ± 2.18.Extreme values were observed at station B4.The average fertility at B4 was 11 ± 2.29 oocytes for a total of 83 ovigerous  females sampled.At station B1, 15 ovigerous females were collected.Absolute fecundity was between 9 and 13 with an average of 10.80 ± 1.32.At station B2, 18 ovigerous females were captured, the minimum and maximum fecundity values being respectively 8 and 15 oocytes (mean= 11.39 ± 2.28).The calculated relative fecundity was 43.52 oocytes per gram of shrimp in all the stations.It was 48.37 ± 6.02, 42.67 ± 8.14 and 42.82 ± 9.18 oocytes per gram at stations B1, B2 and B4, respectively.The analysis of variance did not indicate any significant spatial variation of the two fecundities (p > 0.05).There was a positive and significant relationship (P<0.05)only between body weight and absolute fecundity (Figure 3C).
The regression equations were Y = 4.69 +0.03X, and r = 0.11 (Figure 3B); Y = 0.906 + 0.007 X and r = 0.10 (Figure 3A) and Y = 4.95 +22.88X and r = 0.43 for fecundity-total length, fecunditycarapace length and fecundity-weight of shrimp relationships, respectively.This means that the number of eggs increases only with increasing weight of the individuals.

First maturity size
The size at first maturity (L50) was estimated for females at each sampling station and in the stream as the standard length at which 50% of the shrimp were mature (Figure 4).For pooled data from all stations, the length at which 50% of females reached maturity was 4.82 mm CL (Figure 4A).The L50 estimated for females   4B), B2 (Figure 4C) and B4 (Figure 4D), respectively.Female individuals from B4 reached sexual maturity before those from the other two stations (B1 and B2).From 10 mm CL, all females were mature (Figure 4).

Oocyte size (diameter and volume)
A total of 1289 eggs were measured.Of these, 526 eggs were stage I, 419 eggs stage II and 344 eggs stage III.The mean oocyte diameter increased from 1.08 ± 0.13 mm at stage I to 1.39 ± 0.20 mm at stage III.The ANOVA test showed a significant difference (F = 33.71,p < 0.001) in egg diameter between maturity stages.Concerning the spatial variation, the average value of the diameter of the oocytes was 1.20 ± 0.30 mm at B1, 1.29 ± 0.28 mm at B2 and 1.18 ± 1.15 mm at B4.As for the seasonal variation in oocyte diameter, the smallest value was obtained in the rainy season (1.19 ± 0.20 mm) and the highest in the dry season (1.26 ± 0.18 mm).No significant variation was observed between stations (ANOVA: F = 2.47, p = 0.089) and between seasons (Student's t test: t = -1.02,p = 0.714).As for egg volume, it did not vary significantly between stations (ANOVA: F = 2.14, p = 0.122) and seasons (Student's t test: t = -0.95,p = 0.52).On the other hand, a significant difference was noted according to the stage of maturity (ANOVA: F = 24.43,p < 0.001).Oocyte volume was between 0.70 ± 0.25 mm 3 (B4) and 0.85 ± 0.39 mm 3 (B2), and between 0.73 ± 0.32 mm 3 (rainy season) and 0.84 ± 0.36 mm 3 (dry season).For the maturity stage, egg volume increased from stage I (0.55 ± 0.16 mm 3 ) to stage III (1.00 ± 0.37 mm 3 ).

Reproductive period
Overall, ovigerous females of D. trispinosa occurred throughout the year (Figure 5).From April to May and from November to December higher percentages (30.43 to 68.12%) of ovigerous females were observed, whereas during other months the percentages of ovigerous females were lower (6.9 to 24.44%).Considering stations, ovigerous females were also sampled every month at B4 (free from any human activity).At this station, the highest percentages of ovigerous females were obtained from May to December with peaks in May (100%) and November-December (75 to 53.85%).At B1, ovigerous females were encountered every month, with the exception of August and September.The percentages were higher in May (33.33%),July (30.77%) and December (40%).Ovigerous females were only captured during five months (May, June, September, November and December) at station B2.

Relationships between environmental and reproductive parameters
The relationships established between environmental and biological variables indicated that only oocyte volume was influenced by temperature and dissolved oxygen (Figure 6).A strong positive regression (r = 0.53, P = 0.012) was recorded between dissolved oxygen and oocyte volume (Figure 6A).This means that the oocyte volume of D. trispinosa increased with the dissolved oxygen rate of the water.In contrast, temperature had a negative influence (r = -0.47,P = 0.025) on oocyte volume of D. trispinosa in Banco stream (Figure 6B).The oocyte volume decreased with the increase of water temperature.

Discussion
Our study provides preliminaries data on them parameter of reproduction of prawn D. trispinosa from watercourse of West Africa.The species reproduction was studied for the first time, giving the opportunity to make useful comparisons with the other small prawn species.The results of this study showed that the spatial variations of the physico-chemical characteristics of waters were marked.Generally, quality water of Banco stream was better upstream than downstream.According to Camara et al. (2009), this could be attributable to several anthropogenic activities such as urban runoff to stream due to direct or unregulated discharges in the Banco stream of sewage water from neighbouring cities.These results w consistent with those of Camara et al. (2012) and Yao et al. (2019) in same stream and those of those of Tchakonté et al. (2014) in five rivers of Douala (Cameroon).The pH values observed along the stream were acidic (5.45 to 6.54, respectively at B4 and B1).The acidity of the Banco stream water would be the consequence of the decomposition of plant organic matter in the water leading to a reduction in dissolved oxygen and an increase in CO 2 in the first layers of the soil.These results corroborate those of Camara et al. (2009) and Yao et al. (2019) in the Banco stream.The runoff water was responsible for the strong conductivities and high turbidity at stations (B2 and B3).The values of the wet bed width and the water depth were considerably higher in the upstream stations than in the downstream stations.Naturally, in a watercourse, these two characteristics generally follow a positive upstream-downstream gradient.These results agree with those of Camara et al. (2009Camara et al. ( , 2012) ) in the Banco stream.
Overall, the sex-ratio was in favour of females whether depending on the stations or the seasons.However, only at station B4, the difference in sex-ratio was significant.This demarcation of the sex-ratio of the balance proportion (1: 1) in favour of the females in habitats free from anthropogenic disturbances, could be due to the fact that this species would naturally be polygamous.Crustaceans exhibit a wide array of mating systems including polygamy (Palaoro & Beermann, 2020).This phenomenon is observed in other crustacean species such as the ectoparasitic crab Dissodactylus primitivus Bouvier, 1917 (Jossart et al., 2014) and the thalassinidean ghost shrimp (Bilodeau et al., 2005), Callichirus islagrande (Schmitt, 1935).Furthermore, among populations of the same species, the sex ratio can also be influenced by environmental conditions, geographical features and anthropogenic interference (Lima et al., 2014).Similar sex ratios have also been observed in Atyidae shrimp of small size (Yam & Dudgeon 2005) and related caridean families (Fransozo et al., 2004;Mantelatto & Barbosa, 2005).According to Rodrigues (1996), the predominance of females in a population, may represent a strategic response to compensate for the low fecundity of the species.D. trispinosa produces few eggs as shown by the results of this study.The absolute fecundity of D. trispinosa varied between 5 and 16 (mean = 11.03 ± 2.18).This fecundity is considered low, indicating that this species has a low reproductive potential.Fecundity is related directly to species life strategy (Oh & Hartnoll, 2004) and is of essential importance in order to estimate the reproductive potential (Figueiredo et al., 2008).In addition, fecundity is a phenotypic characteristic that is affected by numerous factors and intensities by specific features of different environments (Hines, 1991) and directly influenced by natural selection (Stearns, 1977).The maximum fecundity of D. trispinosa was much lower than that of other small shrimp species as Caridina cantonensis Yü, 1938, Caridina fossarum Heller, 1862, and Caridina serrata Stimpson, 1860.The results also showed that the increase in the mean egg number follows the increase in individual's weight.Therefore, the weight was an important biological factor which affects the potential fecundity of D. trispinosa.Lambert, (2008) reported that animal fecundity varies with body length or weight.Similar results were observed by Meireles et al. (2013) in Macrobrachium amazonicum from Santa Barbara region (Brasilia).(.No significant relationship observed between the number of oocytes and the size of the shrimp (total length and carapace length), contrary to results observed in shrimp species such as Caridina sakishimensis (Fujino and Shokita, 1976) and C. typus (H. Millne Edwards, 1937) by Soomro et al. (2011), and C. fossarum by Zare et al. (2011).
The occurrence of ovigerous females in all monthly samples in station free from human activities indicates that the D. trispinosa population from Banco stream reproduces continuously year-round.The greater percentage of ovigerous females observed in April, may, November and December which are rainy season months in the study area, suggests that the reproductive cycle of D. trispinosa from Banco River was at its height during rainy periods.Because no partial spawnings were observed, it is assumed that spawnings in D. trispinosa are total.According to Galvão & Bueno (2000) and Mossolin & Bueno (2002), other caridean shrimps, such as the Atyidae Atya scabra (Leach, 1816) and the Palaemonidae Macrobrachium olfersi (Wiegmann, 1836) follow a similar reproductive pattern as that reported here for D. trispinosa.Regarding the breeding period, ovigerous females of D. trispinosa were obtained in all months of sampling.This means that this shrimp has a continuous reproductive period.Similar results have been found by Hoffmann & Fransozo (2010) in the Atyidae shrimp Potimirim glabra (Kingsley, 1878).The first maturity size obtained at stations B1, B2, and B4 was 4.97, 5.70 and 4.23 mm CL, respectively.Which clearly suggested that D. trispinosa quickly reach maturity in station free from anthropogenic disturbances (B4) compared to stations that received them (B1 and B2).This could mean that habitat degradation is responsible for the delay in sexual maturity in this shrimp species.Oocyte size and oocyte volume increase during the incubation period in D. trispinosa.This phenomen is a result of gradual water uptake during embryogenesis in decapods (Müller et al., 2004).According to Hernáez & Palma (2003), the swelling of the egg permits the development of the embryo which is probably controlled by the thickness and permeability of the egg membrane.Similar observations were made in caridean shrimps such as Macrobrachium olfersi by Mossolin & Bueno (2002), M. potiuna (Müller, 1880) by Nazari et al. (2003) and M. carcinus (Linnaeus, 1758) by Lara & Wehrtmann (2009).
The analysis of relationship between physicalchemical variables and reproductive parameters indicated that only temperature had a negative influence on oocyte volume compared to dissolved oxygen.This result suggested that conversion efficiency of yolk reserves in developing embryos in D. trispinosa was significantly reduced at elevated temperature as showed by Brillon et al. (2005) in the shrimp Pandalus borealis Krøyer, 1838.According to Brillon et al. (2005), ovigerous females at high temperature had a lower energetic condition than females at low temperature.At high temperature, ovigerous females accumulated less lipid reserves than ovigerous females at low temperatures.Lower lipid reserves could indicate that higher metabolic costs are associated with respiration and parental care in ovigerous females held at high temperature (Brillon et al. (2005).
In conclusion, the analysed environmental parameters highlighted two groups of habitats: disturbed and undisturbed.The shrimp D. trispinosa reproduces year-round in undisturbed stations and reaches sexual maturity faster at these stations compared to disturbed stations.The sex ratio was in favour of females.This species has a low fecundity (5-16 eggs).
obtained by direct counting of the eggs of 116 ovigerous females taking into

Figure 1 .
Figure 1.Location of sampling stations (B1 to B4) on the Banco stream.

Figure 2 .
Figure 2. Ordination of the physico-chemical variables of the Banco stream sampling stations based on a Principal Component Analysis (A = histogram of eigenvalues; B = correlation circle; C = factorial map); temp = temperature; DO = dissolved oxygen; Width = width of the wet bed; Turb = turbidity; Cond = conductivity.

Figure 3 .
Figure 3. Relationships between absolute fecundity and carapace length (A), total length (B) and individual weight (C) in Desmocaris trispinosa from the Banco stream.
was 4.97, 5.70 and 4.23 mm LC at stations B1 (Figure

Figure 4 .
Figure 4. Logistic curve of estimated sexual maturity for Desmocaris trispinosa in Banco stream.Dashed line represents the class interval where the probability of being mature is 50%: (A) Combined populations; (B) Station B1; (C) Station B2; and (D) Station B4.

Figure 5 .
Figure 5. Monthly changes in percentage occurrence of ovigerous females of Desmocaris trispinosa in Banco stream during july 2021 to June 2022.

Figure 6 .
Figure 6.Relationships between oocyte volume and dissolved oxygen (A) and temperature (B) in Desmocaris trispinosa from the Banco stream.

Table 1 .
Spatial variation of chi-square values for the sex-ratio of Desmocaris trispinosa from Banco stream.

Table 2 .
Seasonal variation of chi-square values for Desmocaris trispinosa sex-ratio Banco stream.