The development of a formulated feed containing Ulva (Chlorophyta) to promote rapid growth and enhanced production of high quality roe in the sea urchin Tripneustes gratilla (Linnaeus)

This study investigated growth and gonad production of Tripneustes gratilla fed four protein-rich artificial diets supplemented with varying amounts (0%, 5%, 15% and 20% weight/weight (w/w); designated 0, 5, 15 and 20U, respectively) of the macroalga Ulva over a 12-week period. Gonad size, texture, colour and a number of production performance parameters were quantified and compared with urchins fed fresh Ulva (FU) and a combination of FU and artificial feed (FB). All artificial diets significantly (P ≤ 0.05) increased gonad somatic indices (GSI) compared with the FU treatment. The 20U treatment increased GSI by 205% by week 9, compared with a 57% increase in the GSI of urchins fed FU. Gonad colour was calculated using three colour parameters, namely L* (lightness), a* (redness) and b* (yellowness). Whereas L* and a* values did not differ significantly between treatments, b* values for all treatments, with the exception of the 20U and FB treatments, were significantly (P ≤ 0.05) lower or less yellow than the FU treatment by week 12. These results show that we have successfully formulated a feed (20U) which can produce commercially acceptable gonads in terms of both size and colour, indicating the potential for this artificial feed to support commercial echinoculture.

Keywords: Echinoculture, Tripneustes gratilla, formulated feed, Ulva, gonad colour, gonad enhancement

 Introduction

The international sea urchin industry relies on the production of marketable quality gonads, which are sold largely for consumption in the sushi restaurant trade. Gonads that are large in size, contain few to no gametes, have a firm texture, and are bright yellow or orange in colour, are regarded as the most commercially valuable (Robinson, Castell & Kennedy 2002; Shpigel, Marciano, McBride & Lupatsch 2004).

Tripneustes gratilla is a fast-growing shallowwater echinoid, occurring throughout the waters of the tropical Indo-Pacific with isolated populations found in warm temperate Indo-Pacific regions (Dworjanyn, Pirozzi & Liu 2007; Rahman, Tsuchiya & Uehara 2009). This species can produce high quality gonads with excellent market acceptance, and is one of the most commercially important sea urchin species in countries, such as Japan (Rahman et al. 2009), where it is regarded, along with tuna, lobster and abalone, as premium seafood. Demand for high quality sea urchin gonads has, however, led to overfishing of natural populations of this and other echinoid species (Rahman et al. 2009). Consequently, many countries are now examining the feasibility of sea urchin aquaculture to supply the demand from future markets (Lawrence, Olave, Otaiza, Lawrence & Bustos 1997; Pearce, Daggett & Robinson 2002a,b; Hammer, Hammer, Watts, Lawrence & Lawrence 2006; Cook & Kelly 2007) and there is also considerable interest in the cultivation of T. gratilla in South Africa, where it occurs naturally on the north-eastern coast. However, in order for echinoculture to be economically viable, high quality feeds need to be formulated to promote fast growth rates and improve the market acceptability of urchin gonads.

Although Tripneustes species are mostly grazers of sea grasses in tropical regions (Dy, Uy & Corrales 2002; Vaitilingon, Rasolofonirina & Jangoux 2003) and macroalgae (Lawrence & Agatsuma 2001) in more temperate regions, such as in South Africa, the use of wild-harvested macroalgae/ sea grass alone as feed for cultured urchins, as well as their inclusion into prepared diets, are unlikely to be commercially viable for large-scale commercial echinoculture. Disadvantages of using naturally harvested macroalgal diets include limited stocks along the Southern African coast; restricted harvesting in certain areas; temporal and seasonal variation in quality and/or quantity; and the negative environmental effects that arise from harvesting large quantities of seaweeds from natural populations (Anderson et al. 2003; Anderson, Rothman, Share & Drummond 2006; Troell, Robertson-Andersson, Anderson, Bolton, Maneveldt, Halling & Probyn 2006; Anderson, Rand, Rothman & Bolton 2007). The primary disadvantage in using natural diets is, however, the relatively low protein content of most naturally occurring seaweeds, with the consequence that they do not support maximal somatic and gonadal growth of sea urchins (Cook & Kelly 2007). For example, the dominant local kelp Ecklonia maxima which is used extensively in abalone culture as a supplement to formulated feeds has a crude protein content of 11–12% of dry weight throughout the year (Smith 2007). In contrast, formulated feeds are considered to be better than natural foods as they can be formulated to a standard composition and promote maximal production (Lawrence et al. 1997). In addition, they are generally considered less expensive, particularly when taking into account differences in food conversion ratios between seaweeds and formulated feeds (Troell et al. 2006).

Over the last decade numerous studies have demonstrated that artificial diets can enhance somatic growth and gonad yield of sea urchins (Lawrence et al. 1997; Olave, Bustos, Lawrence & Carcamo 2001; Shpigel, McBride, Marciano, Ron & Ben-Amotz 2005), and protein has been identified as the main contributing factor for this enhancement (Pearce et al. 2002a,b; Pearce, Daggett & Robinson 2004). In nature, sea urchins must ingest and process large quantities of macroalgae and/or sea grass to meet their nutritional requirements for protein (Hammer et al. 2006), possibly explaining why administration of macroalgal diets alone typically produces urchins with lower growth rates and smaller sized gonads when compared with urchins fed diets formulated using animal-derived material (Cook, Kelly & McKenzie 1998; Fernandez & Boudouresque 1998, 2000). Hammer et al. (2006) demonstrated that an artificial diet containing a moderate protein level of approximately 20% is used most efficiently by sea urchins, resulting in enhanced consumption rates, survival, specific growth rates and gonad production efficiency, when compared with a diet containing low levels of protein (<9%). These findings are supported by several other studies, which also indicate that a moderate dietary protein level of approximately 20% is most efficiently utilized by sea urchins (Akiyama, Unuma & Yamamota 2001; Pearce et al. 2002a; Schlosser, Lupatsch, Lawrence, Lawrence & Shpigel 2005).

Other components of an artificial feed which affect gonad yield and quality include pigment concentration (especially carotenoids), feed stimulants, feed binder type and feed shape (Plank, Lawrence, Lawrence & Olvera 2002; Pearce et al. 2004; Hammer et al. 2006; Cook & Kelly 2007). Inclusion of small quantities of macroalgae, particularly the brown fucoid seaweed Sargassum linearifolium, in artificial diets fed to T. gratilla has been shown to enhance palatability (Dworjanyn et al. 2007). Carotenoid pigments, mainly β-carotene , which is synthesized by macroalgae, are the source of the red, orange or yellow gonad colouration of sea urchin gonads (Agatsuma, Sato & Taniguchi 2005), and addition of β-carotene to artificial feeds has been shown to improve gonad colour of cultured urchins (Pearce et al. 2002a; Robinson et al. 2002; McBride, Price, Tom, Lawrence & Lawrence 2004; Shpigel et al. 2005). However, despite these potential benefits, the effects of a protein-rich formulated feed supplemented with varying amounts of a particular macroalga on the somatic and gonadal growth of T. gratilla have not yet been investigated. A preliminary study demonstrated that T. gratilla consumed significantly higher amounts of Ulva rigida when offered four seaweed species [Ulva rigida, Ecklonia maxima (kelp), Porphyra capensis and Gigartina polycarpa] in paired consumption tests (Scholtz 2008). As South African Tripneustes generally prefer Ulva and because farm-grown Ulva is available in large quantities from local aquaculture facilities (Bolton, Robertson-Andersson, Shuuluka & Kandjengo 2009), Ulva was selected as an additive for an experimentally formulated urchin feed.

The aim of this study was to determine the effects of incorporating varying levels of the algae Ulva into an artificially formulated feed, with the specific aim of increasing gonad mass and improving gonad colour and quality of the sea urchin Tripneustes gratilla. Ulva was incorporated into the formulated diets to act as both a natural feeding stimulant (Dworjanyn et al. 2007) as well as a source of β-carotene ; an important carotenoid pigment associated with gonad colouration. Ulva lactuca has been shown to contain between 25 and 45 μg g ^-1 [fresh weight (FW)] of β-carotene (Bischof, Hanelt & Wiencke 2002), whereas the total carotenoid content of farmed Ulva rigida and Ulva lactuca has been reported as high as 10% [dry weight (DW)] (Shuuluka 2011). Urchins fed the various diets were then assessed on a monthly basis to record somatic growth, gonadal growth and a number of gonad quality factors, such as colour, texture, gonad somatic index and the dominant maturity stage of urchins in each experimental group. 

Adult Tripneustes gratilla (50–70 mm test diameter) were collected during low tide from shallow rock pools near Haga-Haga, Eastern Cape, South Africa (32°45′4.23″S, 28°16′41.30″E) on 15–17 September 2008. Haga-Haga has a warm temperate marine regime (Stephenson 1948), with a mean annual seawater temperature of 17.7°C, and the mean temperature for the coldest and warmest months of the year being 15.8 and 21°C respectively (Bolton 1986). These temperatures are at the lower limit of this species’ habitat temperature, which ranges from 18 to 29°C (Rahman et al. 2009), with T. gratilla larvae, juveniles and adults typically being cultured at a temperature of 24–25°C (Dworjanyn et al. 2007; Dworjanyn & Pirozzi 2008). Following collection, animals were transferred to the Department of Agriculture, Forestry and Fisheries (DAFF) Aquaculture Research Facility in Cape Town, South Africa. Sea urchins at this facility were held in plastic crates (L9W9H: 60940920 cm; with W9L: 3930 mm slits along the sides and bottom) suspended in four large fibreglass tanks (L9W9H: 282 9 182950 cm) supplied with heated recirculating seawater maintained at a salinity of 35 and 24– 25°C under fluorescent lights set to provide a 12:12 day. Seawater flowed through each tank at a rate of 10 L min ^-1 and the tanks were constantly aerated. The recirculating seawater systems were also equipped with a common sand filter, bio-filter and protein skimmer to maintain optimal water quality conditions. Dissolved oxygen (>90%) and pH (8.0 ± 0.2, mean ± SE) were recorded weekly using a Cyberscan PD300 waterproof hand-held pH and dissolved oxygen metre (Eutech Instruments Pte/Oakton Instruments, Vernon Hills, IL, USA), whereas temperature and salinity were recorded using a hand-held WTW LF340 salinity and temperature metre (Wissenschaftlich- Technische Werksta¨tten, Weilheim, Germany). Ammonia and nitrates were measured weekly using Sera® ammonium/ammoniak- and nitrit-test kits, respectively, and values remained within the parameters for a healthy recirculation system (<0.003 and <0.001 mg L ^-1 for ammonia and nitrate respectively). To reduce differences in gonad development and standardize nutritional condition, sea urchins were acclimated for at least 3 months and fed a diet of fresh kelp, Ecklonia maxima, every second day. Before each feeding, uneaten food was removed and fresh kelp was then replenished.

The effects of six dietary treatments on gonad production and quality factors, such as colour, texture and firmness, of adult T. gratilla were assessed in a 12-week feeding trial. Four of the dietary treatments consisted of a semi-purified (contain natural ingredients in as pure a form as is available) ‘basal’ formula supplemented with different amounts of dried macroalga Ulva spp. to achieve final concentrations of 0, 5, 15 and 20% (w/w). The proximate nutrient analysis of the six dietary treatments is presented in Table 1. The Ulva used in this study was produced in paddle raceway aquaculture systems at Irvine & Johnson (I&J) Cape Abalone farm (34o34′60 S; 19o21′0 E) and consisted of a mixture of predominantly Ulva rigida with some ‘Ulva lactuca’ (Robertson-Andersson, Potgieter, Hansen, Bolton, Troell, Anderson, Halling & Probyn 2008; Bolton et al. 2009; Shuuluka 2011). South African U. lactuca (sensu Stegenga et al. 1997) is, however, not the same species as the U. lactuca known elsewhere in the world, and currently has no valid name (L. Kandjengo, University of Namibia, pers. comm.,). Ulva collected from this facility was dried for several days in ovens at 60°C before being ground to a fine powder using a hammer mill. The basal diets supplemented with and without Ulva were formulated and manufactured at the Division of Aquaculture, University of Stellenbosch, South Africa. Diets were produced in the form of semi-moist extruded chips (L 9 W: 2 9 4 cm), which were dried to a constant weight in a drying oven at 70°C. To ensure that the diets remained functionally equivalent, an artificial Ulva additive (designed to have the same nutritional properties as Ulva; Table 1) was created using different dietary ingredients (Table 2), so that varying amounts of Ulva spp. could be substituted out of the treatments without significantly changing the nutritional value and characteristics of the feed. This was done by conducting a complete nutrient analysis (proteins, carbohydrates, amino acids etc.) on the Ulva collected from I&J (Table 3). In addition to the extruded feeds, fresh cultivated Ulva was included as a dietary treatment to act as a control to determine the effects of a fresh macroalgal diet alone on gonad yield and quality. A mixed diet of fresh Ulva and basal feed (FB), the formula used to make the artificial diets prior to the addition of Ulva or Ulva additive, was also included to determine the effect of using fresh Ulva compared with dried Ulva on gonad yield and characteristics of urchins fed artificial feed. The treatments were abbreviated as follows: 20% Ulva = 20U; 15% Ulva = 15U; 5% Ulva = 5U; 0% Ulva = 0U; Fresh Ulva = FU and fresh Ulva and the basal diet = FB. 

Table 1 Proximate nutrient analysis of fresh Ulva, the Ulva additive, the basal formula (no Ulva or Ulva additive included) and the four prepared diets fed to Tripneustes gratilla during the experiments (per g dry matter). Treatment groups: FU = fresh Ulva; 20U = 20% Ulva; 15U = 15% Ulva; 5U = 5% Ulva; 0U = 0% Ulva

 

Table 2 Ingredients and the ratio at which they were combined to formulate the Ulva additive that was incorporated into the experimental diets

 

Table 3 Proximate nutrient composition (per g dry matter) of the Ulva collected from Irvine & Johnson (I&J) Cape Abalone farm in the Western Cape Province of South Africa for the preparation of the experimental diets utilized in this study

Nutritional analysis was conducted at the Department of Food and Animal Science, University of Stellenbosch, to determine whether the formulated diets were similar in nutrition, even though they contained varying amounts of Ulva and/or Ulva additive. The four artificial diets as well as Ulva, Ulva additive and the ‘basal’ formula, were analysed to determine their crude protein, fat, moisture, ash, fibre, carbohydrate and gross energy contents using protocols described by AOAC International (2002). The results of the nutrient analysis are presented in Tables 1 and 3. 

The feeding trial was initiated on the 11 March 2009 and ran for a total of 90 days, concluding on the 8 June 2009. A total of 12 baskets were suspended in each of the four large fibreglass tanks and were stocked at an initial stocking density of 15 animals/basket. Within each raceway there were two replicates of each treatment, resulting in an overall experimental design that consisted of eight replicates and six treatments (20U, 15U, 5U, 0U, FU and FB). These treatments were randomly allocated to baskets within each of the four fibreglass tanks at the start of the feeding trial to account for the potential effects of specific feeding regimes on surrounding baskets within a tank. Sea urchins were fed each experimental diet ad libitum every second day after the removal of uneaten feed from each basket.

Immediately before the start of the experiment, 10 sea urchins were randomly selected and dissected to establish the initial state of the animals. Thereafter, one urchin from each basket was sampled at random each month for the duration of the feeding trial. Each sampled urchin was blotted dry with paper towel, and total body weight, test diameter, test height and urchin drained weight (coelomic fluid removed) were carefully recorded to the nearest 0.01 g or 0.01 mm. The gonads were then carefully dissected out and the gonad wet weight recorded to the nearest 0.01 g using an electronic balance. Gonad texture, firmness and colour were rated visually and manually by a single observer according to the procedures described below. A separate gonad from each urchin was transferred into Davidson’s Fixative (per litre: 300 mL 95% ethyl alcohol, 200 mL 100% formalin, 100 mL glycerol, 100 mL glacial acetic acid and 300 mL distilled water) immediately following dissection and fixed for 48 h before being transferred into 70% ethanol and processed for routine paraffin histology (Bucke 1989).

The body and gonad wet weights of individual urchins fed the various experimental diets (n = 8) were used to calculate gonad somatic index according to the following formula used by Pearce et al. (2002a):

GSI(%) = W_g/W_t x 100;

where W_g is the wet weight (g) of the gonad and W_t is the total wet weight (g) of the sea urchin.

To allow us to compare the growth rates of urchins fed formulated feeds in this study to other similar studies, we calculated the GSI increase per week over the course of the study using the following formula:

Gonad yield per a week = [(GSI_(end) - GSI_(start)) / (#days)] x7;

where GSI_(end) is the GSI at the end of the study and GSI_(start) is the GSI at the start of the study.

Eye rated gonad colour was assessed by visually ranking each gonad (n = 8 per treatment) in categories ranging from most desirable to unacceptable according to the rating system described by Pearce et al. (2002a). These observations were made by a single observer [Mark D. Cyrus (MDC)] over the course of the study. The categories were allocated numbers which were ranked as follows: (1) bright yellow-orange gonads (excellent quality); (2) yellow- orange gonads (acceptable quality); (3) pale yellow-orange or dark yellow-orange gonads (low quality) and (4) white or brown gonads (unacceptable). Values from within the different dietary treatments were averaged to produce a single value that could be compared between treatments. The lower the average eye rated colour value, the better the colour and therefore quality of the gonads.

Spectrophotometer rated gonad colour was objectively quantified using a hand-held reflected-light, fibre-optic spectrophotometer [Gardner (BYK) Colour Guide, Wesel, Germany]. Three replicate measurements of intensity of lightness (L*), redness (a*) and yellowness (b*) for each gonad were recorded. These values were then used to determine the average L*, a* and b* values for each measured gonad sample. Total difference in gonad colour from A grade roe (the fresh Ulva treatment which was regarded in this study as having the most desirable gonad colour) was calculated for urchins fed the six diets by comparing gonad colour in these treatments to the average colour of urchin gonads fed the fresh Ulva spp. The following formula from McBride et al. (2004) was used to calculate these differences:

ΑE_ab*=[(L* _Ulva - L* _Sample)² + (a* _Ulva - a* _Sample)² + (b* _Ulva - b* _Sample)²]½

Texture affects gonad appearance and was evaluated according to the visibility of individual gonad follicles (smoothness) and segment halves, where gonads with greater separation between follicles represented a lower quality product. Individual gonads (n = 8 per treatment) were visually assessed by a single observer (MDC) using the protocol from Pearce et al. (2002a) and were allocated a rating as follows: (1) two distinct gonad segment halves with little to no follicle separation; (2) two distinct gonad segment halves with visible follicle separation; (3) distinction of gonad segment halves possible, but quite granular and (4) distinction of gonad segment halves not possible, gonads rough/granular.

The textural properties and hence the firmness of gonads is very important as gonads that remain intact during processing and packaging are required to produce a high quality marketable product. Firmness is defined as the force required to compress a sea urchin gonad, a fixed distance for a specific time, however, because we lacked the equipment to test firmness, gonads (n = 8 per treatment) were visually rated subjectively by pressing down on the gonad with a finger and scoring each gonad according to one of the following subjective criteria described by Pearce et al. (2002a): (1) very firm; (2) firm; (3) soft and (4) very soft. These observations were made by a single observer (MDC) over the course of the study.

Histological analysis was conducted to determine the amount of gametogenic activity in the gonads of urchins fed the various diets, with high quality sea urchins gonads considered to be those which contain little to no gametogenic activity. Fixed tissues (n = 8/treatment/sampling date) were processed using standard histological techniques (Bucke 1989) and the paraffin wax embedded samples were sectioned on a microtome at 3 lm. Histological sections were stained with Harris’s haematoxylin and eosin and examined under an Olympus BX 51 light microscope equipped with a Leica digital camera and Nikon Imaging Systems (NIS) Elements Basic Research (BR) image analysis software (Version 3.1). Individual gonads were categorized into one of six different maturity stages according to published literature on echinoid gametogenesis (Byrne 1990; Spirlet, Grosjean & Jangoux 1998). These maturity stages were based on oocyte size (females), thickness of peripheral spermatocyte layer (males) and the amount of non-germinal nutritive tissue (males and females), and categorized as follows: (1) recovery; (2) growing; (3) premature; (4) mature; (5) partly spent and (6) spent. Of the six categories, gonads in the recovery and growing phases were regarded as high quality gonads with good market acceptance.

Lastly, the density of nutritive phagocytes (NP) within each gonad was calculated to determine whether the density of NP within each gonad differed between the dietary treatment groups or stages of gonad maturity. Stained histological sections of gonads were examined under an Olympus BX 51 light microscope and photographed using a Leica digital camera. Nikon Imaging Systems (NIS), Elements Basic Research (BR) image analysis software (Version 3.1) was used for all image analysis. Three randomly allocated regions of interest (ROI) were selected from each image using Pixel Classifier software, which accurately separates dyed areas within the gonad occupied by NP’s, gametes or unoccupied space, making it possible to quantify the relative area that each of these three cell types occupy within the ROI. The classifier allows for segmentation of the image pixels according to different user-defined classes and is based on pixel features ,such as intensity values and RGB values. The Pixel Classifier software accurately records the area occupied by NP’s, gametes and unoccupied space which can then be used to determine NP density. The values obtained from each of the three ROI were then averaged to get a mean ROI for each image (Fig. 1a–d). 

Figure 1 Histological section of gonad from Tripneustes gratilla fed either fresh Ulva (a and b) or the 0% Ulva supplemented diet (0U treatment, c and d). Images a and c represent the original images (4009 magnification), whereas the pixels within images b and d have been subdivided based on pixel intensity values and RGB values, allowing for the separation of areas within the gonad occupied by nutritive phagocytes (NP), gametes or unoccupied space. Sections shown the accumulation of NP’s within gonads of urchins fed fresh Ulva (a and b) compared with the 0U treatment group (c and d).

 Statistical analysis

To determine whether urchin wet weight, mortality, GSI, gonad wet weight and gonad colour (L*, a* and b*) changed as a function of time within individual treatment groups or as a function of treatment at individual sampling dates, a one-way analysis of variance (ANOVA) was performed using Statistica 8 statistical software. All tests for normality (Kolmogorov–Smirnov test) and equal variance (Levene’s test) passed for all data sets. Oneway ANOVA was also used to test for significant differences in eye rated gonad colour, gonad texture and firmness and gonad maturity (histological data) within individual treatment groups over time and between treatment groups at specific sampling dates. All tests for normality (Kolmogorov–Smirnov test) and equal variance (Levene’s test) in these data sets failed and therefore a Kruskal–Wallis ANOVA on Ranks test was performed to test for significant differences. The Tukey method was used for all post-hoc multiple comparisons between individual timepoints within a treatment group and between the different treatment groups at specific sampling dates. Significance was assigned to P-values of <0.05 for all analysis.

A Fit Generalized Estimating Equations General Linear Model (geeglm) was used to determine variation in the spectrophotometer rated gonad colour values (L*, a*, b*) in relation to diet type, sex of urchin, maturity level and GSI using the R package geepack for generalized estimating equations (Halekoh, Højsgaard & Yan 2006). Unlike the General Linear Model, where observations are assumed to be independent, Generalized Estimating Equations (GEE) are particularly useful for observations which are clustered or dependent on one another, such as the image data recorded in this study. All tests for normality and equal variance passed for all data sets. Each geeglm consisted of diet type, sex of urchin, maturity level and GSI as independent variables and the spectrophotometer rated gonad colour values (L*, a*, b*) as dependent variables. All possible combinations between the independent variables were treated as interaction effects for this analysis. A Generalized Linear Model (GLM) was used to determine variation in the density of nutritive phagocytes within each gonad in relation to wet gonad weight, diet type, sampling date, gonad maturity and sex of the urchin. For this GLM, wet gonad weight, diet type, sampling date, gonad maturity and sex of the urchin were included as independent variables, and the density of nutritive phagocytes as the dependent variable. As before, all possible combinations between the independent variables were treated as interaction effects for this analysis. Significance was assigned to P-values of <0.05 for all analysis. 

Tripneustes gratilla survival rates over the course of the study were high and did not vary significantly between diets, ranging from 92.5% for animals fed the 0U diet to 97.5% for animals fed fresh Ulva (FU, data not shown).

Somatic growth, determined using urchin wet weight, increased significantly (one-way ANOVA, P ≤0.05) within all dietary treatment groups over the 12-week experimental period (Fig. 2a). Overall, the mean wet weight (±SE) of urchins across all treatment groups increased from 99.08 ± 10.15 g at the beginning of the study to 163.13 ± 4.07 g at the end of the study. Urchins fed the artificial diets showed an average gain in wet weight of 68.2% over the course of the 12- week study compared with a 46.7% increase in wet weight for urchins fed fresh Ulva. There was, however, no significant difference (one-way ANOVA, F_5,42 = 0.813, P = 0.546 for the 12-week sampling date) in urchin wet weight between any of the treatment groups by the end of the study period. Similar results were recorded for urchin test diameter (data not shown).

Urchins fed the formulated feeds had a significantly greater gonad wet weight and GSI by week 9 and at the end of the experiment when compared with urchins fed exclusively on a diet of fresh Ulva (one-way ANOVA; P≤0.001 for gonad wet weight and GSI for the 12-week sampling date) (Fig. 2b and c). By week 9, urchins fed formulated feeds had already achieved a 190 ± 12.09% (mean ± SE) increase in GSI, from the initial starting value, compared with a 57.3% increase in GSI for animals fed exclusively with fresh Ulva over the same time period (Fig. 2c). There was, however, no significant difference between the GSI of urchins fed the different formulated feeds at any stage of the experiment. 

Figure 2 Mean (a) urchin wet weight, (b) gonad wet weight and (c) gonad somatic index (GSI) of Tripneustes gratilla fed with four artificial diets, fresh Ulva or a mixed diet consisting of fresh Ulva and a basal feed over a 12-week grow-out period in recirculating seawater systems. Data represents mean ± SE of eight replicates per treatment group. *(P < 0.05, Tukey test) represents a significant difference in the means of urchins fed the artificial or mixed diets from the means of urchins fed fresh Ulva. Treatment groups: FU = fresh Ulva; FB = fresh Ulva and the basal diet; 20U = 20% Ulva; 15U = 15% Ulva; 5U = 5% Ulva; 0U = 0% Ulva.

Mean eye rated gonad colour (Fig 3) ranged from 1.8 to 3.4 for urchins fed the formulated feeds, whereas urchins fed with fresh Ulva produced lower colour ratings which fluctuated between 1.4 and 2. At the end of the feeding trial, urchins fed the 20U diet and fresh Ulva had comparable colour ratings that were not significantly different from one another (Kruskal–Wallis, P > 0.05). In contrast, the gonad colour of urchins fed the remaining three artificial diets (0U, 5U, 15U) was significantly higher, or worse in terms of market acceptance, than the fresh Ulva treatment (Kruskal –Wallis, H₅ = 22.37, P = 0.0004 for the 12-week sampling date) by the end of this study. 

Figure 3 Mean eye rated (ER) gonad colour of Tripneustes gratilla fed with four artificial diets, fresh Ulva or a mixed diet consisting of fresh Ulva and a basal feed over a 12-week grow-out period. Data represents the mean ± SE of eight replicates per treatment group. Gonad colour was visually ranked in categories ranging from most desirable to unacceptable. The categories were allocated numbers, which were ranked as follows: (1) bright yellow-orange gonads (excellent quality); (2) yellow-orange gonads (acceptable quality); (3) pale yellow-orange or dark yelloworange gonads (low quality) and (4) white or brown gonads (unacceptable). Treatment groups: FU = fresh Ulva; FB = fresh Ulva and the basal diet; 20U = 20% Ulva; 15U = 15% Ulva; 5U = 5% Ulva; 0U = 0% Ulva.

Gonad lightness (L*) values ranged from 57.88 to 62.44 and were unaffected by diet, with no significant difference recorded between any of the dietary treatments by the end of this study (Fig. 4a). Likewise, gonad redness (a*) values did not vary significantly between treatment groups over the course of the study and ranged from 11.17 to 15.22 (Fig. 4b). In contrast, mean gonad yellowness (b*) values decreased with the decreasing Ulva content of each diet and by week 9 urchins in the 5U treatment group showed significantly lower gonad b* values when compared with the FU treatment group (Fig. 4c; one-way ANOVA; F5,42 = 3.891, P = 0.001). By the end of the feeding trial, the 15U, 5U and 0U treatment groups had significantly lower b* values (F_5,42 = 5.92, P = 0.0003) compared with the FU treatment group. In contrast, urchins in the 20U and FB treatment groups produce gonads with similar yellowness values that were not significantly different from urchins fed a diet of fresh Ulva by the end of the study. The total difference in gonad colour, calculated using the formula from McBride et al. (2004), between urchins fed fresh Ulva and the various artificial diets support the latter findings, with total gonad colour values increasing as the Ulva content of each diet decreased (Fig. 5). The 20U and FB treatment groups did, however, produce urchins with gonads that were similar in colour to urchins fed fresh Ulva, and these two treatments had significantly better colouration (one-way ANOVA, F = 2.901, d.f. = 39, P = 0.035 for the 12-week sampling date), in terms of their market acceptance, compared with the 15U, 5U and 0U treatment groups by week 12.

In addition, a Fitted Generalized Estimating Equations General Linear Model (geeglm) revealed that gonad colour (L*, a*, b*) was significantly influenced by dietary treatment and gonad maturity, as gonads in the FU treatment group were significantly better coloured, in terms of their market acceptance, than those in the 15U (P = 0.016), 5U (P = 3.1e-05) and 0U (P = 0.0001) treatment groups (Table 4). The 20U and FB treatment groups, however, did not vary significantly in terms of gonad colour when compared with the FU diet. Gonad maturity also contributed to and significantly affected gonad colour, with gonads in a recovering stage having similar coloured gonads to those in a spent stage. Whereas gonads in the growing (P = 0.0071), pre-mature (P = 0.0182), mature (P = 0.0005) or partly spent (P = 0.0016) states had significantly better coloured gonads than those in a recovery or spent state (Table 4). 

Figure 4 Mean gonad (a) L* (lightness), (b) a* (redness) and (c) b* (yellowness) values of Tripneustes gratilla fed with four artificial diets, fresh Ulva or a mixed diet consisting of fresh Ulva and a basal feed over a 12- week grow-out period. Data represent the mean ± SE of eight replicates per treatment group. *(P < 0.05, Tukey test) represents a significance differences in the means of urchins fed the artificial or mixed diets from the means of urchins fed fresh Ulva. Treatment groups: FU = fresh Ulva; FB = fresh Ulva and the basal diet; 20U = 20% Ulva; 15U = 15% Ulva; 5U = 5% Ulva; 0U = 0% Ulva.

 

Figure 5 The total difference in spectrophotometer rated gonad colour of Tripneustes gratilla fed with the four artificial diets or a mixed diet (consisting of fresh Ulva and a basal diet) from the gonad colour of urchins fed with fresh Ulva. Data represent the mean ± SE of eight replicates per treatment group. Treatment groups: FU = fresh Ulva; FB = fresh Ulva and the basal diet; 20U = 20% Ulva; 15U = 15% Ulva; 5U = 5% Ulva; 0U = 0% Ulva.

 

Mean gonad texture ratings at the end of the study period ranged from 1.75 to 2.88 and all treatment groups consisted of gonads with two distinct gonad segment halves with at least some visible follicle separation. There were, however, no significant differences in gonad texture ratings between any of the treatment groups at each of the sampling dates. Similarly, mean gonad firmness ratings did not differ significantly between treatments at the end of the study period and ranged from 2 to 3, indicating that gonads were soft– firm in their appearance.

Histological analysis of sea urchin gonads revealed a significant degree of variability in overall gametogenic development, and therefore gonad maturity, between sampling dates within all treatment groups (Kruskal–Wallis, H = 72.68, d.f. = 3, P = 6.565e-007), however, there were no significant differences between the different dietary treatment groups at each sampling date (Fig. 6). At the beginning of the feeding trial, the gonads of approximately 50% of urchins were in a recovery stage, whereas the remaining 50% were split evenly with 25% being partly spent and 25% being completely spent (Fig. 6a). Thereafter, there was a fairly consistent progression in gonad maturity, with gonads developing or maturing throughout the study period and accumulating nutritive phagocytes, which were stored and then used to produce gametes. 

Table 4 A Fit Generalized Estimating Equations General Linear model (geeglm) analysis showing the significant interactions between dietary treatment, sex of an urchin, certain maturity stages and Gonad Somatic Index (GSI) when predicting spectrophotometer rated gonad colour values (L*, a*, b*) of Tripneustes gratilla fed with four artificial diets, fresh Ulva and a mixed diet over a 12-week grow-out period. All non-significant interactions were excluded from the table. Treatment groups: FU = fresh Ulva; FB = fresh Ulva and the basal diet; 20U = 20% Ulva; 15U = 15% Ulva; 5U = 5% Ulva; 0U = 0% Ulva

Although gonad maturity did not differ significantly between the treatment groups, a General Linear Model analysis revealed (Table 5) that the density of NP is significantly affected by dietary treatment, gonad maturity and sex of the urchin. During week 9, diet type significantly (F5,37 =2.743, P = 0.033) affected the storage of nutrients and therefore the density of NP’s (Fig. 7) within the gonad. Urchins in the 20U treatment group had significantly more densely packed NPs compared with the 5U (P = 0.005) and 0U (P = 0.025) treatment groups, tending to indicate that the addition of Ulva to prepared diets may affect cell density. Gonad maturity was also shown to have an effect on NP density with gonads in the recovery stage having significantly denser packed NP’s compared with gonads in a partly spent (P = 0.006) or spent stage (P = 0.007) (Table 5). Sex was also shown to significantly (P < 0.00001) affect NP density within the gonad as female gonads had more densely packed NP’s compared with male gonads. Nutritive phagocytes density therefore appeared to show a general trend of decreasing as gonads developed gametes and the gonads mature. 

The artificial diets formulated and tested in this study have been shown to significantly increase gonad yield in the sea urchin Tripneustes gratilla when compared with urchins fed a diet of fresh Ulva. The GSI of urchins fed artificial diets in this study ranged from 20.05% to 24.96% (≈200% increase) at the end of the 12-week study, whereas urchins fed exclusively with a natural algae diet had significantly lower GSI values, reaching a maximum of just 12.5% (64% increase) within the same time period. These findings are consistent with the results of previous studies that have demonstrated the effectiveness of artificial feeds for enhancing the growth and development of sea urchins, with artificial diets having successfully increased the GSI of cultured Evechinus chloroticus (James 2006; Phillips, Hamid, Silcock, Sewell, Barker, Weaver, Delahunty & Bremer 2009; Phillips, Bremer, Silcock, Hamid, Delahunty, Barker & Kissick 2010), Psammechinus miliaris (Pantazis, Kelly, Connolly & Black 2000), Paracentrotus lividus (Spirlet, Grosjean & Jangoux 2000; McBride et al. 2004; Schlosser et al. 2005; Shpigel et al. 2005), Strongylocentrotus droebachiensis (Walker & Lesser 1998; Pearce et al. 2002a, 2004), Lytechinus variegatus (Hammer, Hammer, Watts, Desmond, Lawrence & Lawrence 2004; Taylor, Powell, Watts & Lawrence 2009) and Loxechinus albus (Lawrence et al. 1997; Olave et al. 2001) when compared with wild caught individuals or individuals fed natural diets. This study also demonstrated that cultured T. gratilla fed ad libitum with the macroalga Ulva achieved significantly higher GSI values (up to 12.5%) when compared with the GSI values (maximum of 4.9%) reported for T. gratilla collected from the wild (Muthiga 2005). However, even though the artificial diets tested in these studies produced large gonads, many of the gonads were pale and unmarketable. In contrast, we demonstrated that artificial diets supplemented with 20% (w/w) dried Ulva (20U) can produce bright yellow, marketable gonads which do not differ significantly in colour in the experimental conditions from those produced on fresh Ulva. 

Figure 6 The mean number of Tripneustes gratilla (n = 8) of both sexes allocated to each maturity stage over the 12-week grow-out period. Gonads of urchins fed the four artificial diets, fresh Ulva and a mixed diet (consisting of fresh Ulva and a basal feed) were processed for routine paraffin histology at (a) week 0, (b) week 3, (c) week 9 and (d) week 12 to assess the amount of gametogenic activity in the gonads and categorized as (1) recovery; (2) growing; (3) premature; (4) mature; (5) partly spent or (6) spent. Treatment groups: FU = fresh Ulva; FB = fresh Ulva and the basal diet; 20U = 20% Ulva; 15U = 15% Ulva; 5U = 5% Ulva; 0U = 0% Ulva.

 

Table 5 A generalized Linear Model (GLM) showing the significant interactions between wet gonad weight, dietary treatment, sampling date, gonad maturity and sex of urchin when predicting the density of nutritive phagocytes within the gonads of Tripneustes gratilla fed with four artificial diets, fresh Ulva and a mixed diet over a 12-week grow-out period. All non-significant interactions were excluded from the table. Treatment groups: FU = fresh Ulva; FB = fresh Ulva and the basal diet; 20U = 20% Ulva; 15U = 15% Ulva; 5U = 5% Ulva; 0U = 0% Ulva

 

Figure 7 Mean nutritive phagocyte density (NP) (percentage ± SE, n = 8) within the gonads of Tripneustes gratilla fed with four artificial diets, fresh Ulva or a mixed diet consisting of fresh Ulva and a basal feed over a 12-week grow-out period. Treatment groups: FU = fresh Ulva; FB = fresh Ulva and the basal diet; 20U = 20% Ulva; 15U = 15% Ulva; 5U = 5% Ulva; 0U = 0% Ulva.

 

Table 6 Gonad growth, expressed as a per cent increase in gonad somatic index (GSI) per week, of a variety of sea urchin species fed with artificial and natural macroalga feeds

The gonad growth rates obtained from urchins fed formulated feeds in this study compare quite favourably to those in other studies in the literature for formulated and natural feeds (Table 6). On average, T. gratilla fed formulated feeds had a 1.2% increase in gonad somatic index per week over the course of this study, which is comparable to the growth rates reported by Pearce et al. (2004) and Shpigel et al. (2005) for Strongylocentrotus droebachiensis and Paracentrotus lividus respectively. Moreover, weekly gonad growth rates on feeds formulated in this study were 19 and 87.5% higher than the growth rates of Evechinus chloroticus in a similar study using two different formulated feeds currently used in commercial echinoculture (Woods, James, Moss, Wright & Siikavuopio 2008). These reports would tend to suggest that some of the feeds currently formulated for commercial echinoculture do not achieve maximum GSI gain and therefore can be improved.

The high GSI values recorded in the literature for urchins fed artificial diets compared with those fed natural algal diets have been attributed, primarily, to differences in protein levels between the feeds. As protein is one of the most expensive components of an artificial feed, it is essential to determine the optimal levels of protein required to maximize growth. Previous studies have determined that a moderate dietary protein level of approximately 20% is most efficient for maximizing the gonad growth of sea urchins (Akiyama et al. 2001; Pearce et al. 2002a; Schlosser et al. 2005). In this study, all artificial diets were formulated to contain roughly equal crude protein concentrations (≈26%), so that the only variable differing between the prepared diets was the inclusion of different amounts of Ulva or Ulva additive. Ulva used in this study had a crude protein content of 18.31%, which is significantly lower than the protein content of the prepared diets (≈26%); possibly explaining the enhanced growth and development of urchins fed prepared diets compared with fresh Ulva. It could, however, also be argued that the enhanced growth of urchins fed the prepared diets in this study may be attributed to differences in consumption rates. Consumption rates for T. gratilla fed the 20U diet were significantly higher compared with animals fed the 5U and 0U diets (data not shown). However, no significant differences in urchin wet weight or gonad wet weight were recorded between these treatment groups (Fig. 2). Also, each urchin consumed, on average, approximately 3.5 g of Ulva (wet weight) per day compared with 1.5–2.0 g of each prepared feed (dry weight) per day (data not shown). As Ulva has a moisture content of approximately 80% (Shuuluka 2011), this equates to approximately 0.7 g of Ulva (dry weight) per urchin per day. These findings suggest that the high protein content and increased consumption of the prepared feeds may have collectively contributed to the enhanced growth of urchins fed the prepared diets in the present study, although other factors, such as feed shape and texture should also be considered. Moreover, our consumption data also suggests that the inclusion of high quantities of dried Ulva into a prepared feed may act as a feeding stimulant.

Schlosser et al. (2005) suggest that digestible energy may also be a limiting factor in naturally available diets, such as Ulva, as they showed that a prepared diet with a crude protein content of 23% produced significantly higher GSI values compared with a natural diet of Ulva lactuca with a crude protein content of 35%. In their study, the prepared feed and fresh Ulva had gross energy values of 19.39 and 13.39 MJ kg ^-1 respectively. Schlosser et al. (2005) believe that the difference in energy content between these two feeds resulted in enhanced growth and gonad development with the prepared feed. The higher energy content of the prepared feed used in that study is similar to the results recorded in the present study, where the prepared feeds in the current study had gross energy contents between 15.49 and 17.18 MJ kg ^-1, whereas the natural Ulva diet had a gross energy content of 9.44 MJ kg ^-1. Both protein and energy contents of Ulva in this study were therefore shown to be suboptimal, as the artificial feeds contained 65% more energy and 44% more protein than Ulva. The low gross energy, crude protein content and consumption of Ulva utilized in this study makes it difficult to determine which of the three factors, if not all, may have contributed to the low GSI values recorded for urchins fed fresh Ulva compared with those fed artificial diets.

Gonads of commercial quality need to be acceptable in terms of colour, which in the past has proven to be more difficult to achieve using artificial diets compared with natural diets (Senaratna, Evans, Southham & Tsvetnenko 2005; Shpigel et al. 2005). Differences in colour between animals fed natural and artificial diets have been attributed mainly to the lack of carotenoid pigments generally contained within artificial diets (Robinson et al. 2002; Shpigel et al. 2005; Shpigel, Schlosser, Ben-Amotz, Lawrence & Lawrence 2006; Symonds, Caris-Veyrat, Kelly & Young 2007). In this study, the addition of Ulva to our formulated diets acted as a natural source of carotenoid pigments, with Ulva rigida and Ulva lactuca having an average total carotenoid content of 6.7% and 7.3% DW respectively (Shuuluka 2011). Echinenone, which is synthesized from β-carotene, is responsible for the yellowish-orange colour of high quality roe (Pearce et al. 2002a; McBride et al. 2004; Shpigel et al. 2005). We demonstrated that gonad colour is dependent on the inclusion ratio (0%, 5%, 15% and 20% w/w) of Ulva within the diets, with higher inclusion rates promoting better coloured gonads. The FU diet, as expected, proved to be the most successful at producing gonads of marketable colour, and this is most likely due to the high concentration of β-carotene contained within this diet compared with the artificial ones. Indeed, the FU diet produced gonads which differed significantly in eye rated colour from all other diets used in this study with the exception of the 20U diet. As eye rated colour may be quite subjective, a non-subjective colour measure was also employed which gave three colour values, namely lightness (L*), redness (a*) and yellowness (b*). In our study there were no significant differences in gonad lightness (L*) or redness (a*) between any of the dietary treatments. However, the yellowness (b*) values recorded in this study varied significantly between the sampling dates and dietary treatments. By the 9th week, FU gonad b* values varied significantly from only the 5U diet. However, by the 12th week the FB and 20U diets did not vary significantly from the FU b* values, whereas diets which contained less Ulva (15U, 5U and 0U diets) produced gonads which were less yellow. These findings reinforce the importance of carotenoid pigments in sea urchin diets for increasing gonad quality and suggest that both the 20U and FB diets can successfully produce gonads that are not significantly different in terms of colour from gonads of urchins fed a natural diet of Ulva. Robinson et al. (2002) reported similar results for gonad yellowness (b*) values during the course of their study, as yellowness decreased significantly over time in gonads of animals fed artificial diets, whereas wild urchins (Strongylocentrotus droebachiensis) feeding on natural algae diets had an increase in gonad yellowness over time. Aside from the obvious commercial interests in higher profits obtained for better coloured gonads, it has been suggested that there are several other advantages to adding carotenoid pigments to prepared sea urchin diets, as pigments within actively growing tissue can act as important antioxidants, are involved in protein stabilization, aid in egg production and provide ultraviolet protection for sensitive tissues (George & Lawrence 2002; Robinson et al. 2002).

In this study, texture and firmness ratings for all treatment groups and sampling dates did not vary significantly over the study period and generally fell within an acceptable range for sale on commercial markets. However, a gonad factor which differed significantly between treatment groups and sampling dates was the density of NPs contained within gonads. There was a trend of decreasing gonad NP volume with decreasing Ulva content, and we demonstrated that the 5U and 0U diets had significantly less densely packed NPs, ranging from 46.8% to 62.2% over the course of the study, compared with the FU, FB, 20U and 15U diets, which ranged from 53.1% to 68.7% and did not differ significantly in NP density from each other. Gonads in a recovery phase were also show to have significantly more densely packed NPs compared with gonads in a partly spent or spent stage, whereas urchins in a growing, premature or mature stage did not differ significantly. This result is, however, not surprising, as development of gonads from recovery to mature stages is characterized by nutrient accumulation and gamete development, whereas immediately before and after spawning nutrients have been used up and most gametes have been released, resulting in a decrease in gonad NP density. Our results support findings by Bo¨ttger, Devin and Walker (2006), who found that S. droebachiensis in early pre-gametogenesis have an increased volume of NP’s within the gonad compared with gonads in later maturity stages. Histological analysis of gonads over the course of this study did, however, reveal no significant difference in gonad maturity between the different dietary treatments. Gonad maturation followed a similar trend in all treatment groups with gonads accumulating nutrients within NPs, followed by the development and storage of gametes.

In conclusion, we have clearly demonstrated that prepared diets (26% crude protein content) can significantly increase gonad yield of urchins within 9 weeks, compared with urchins fed only a natural diet of fresh Ulva. More importantly, we have demonstrated that when these artificial diets are supplemented with 20% (w/w) dried Ulva, marketable quality gonads can be produced which do not differ significantly in colouration from the gonads of urchins fed with fresh seaweed. Therefore, there is clearly potential for artificial diets containing Ulva to support the commercial cultivation of Tripneustes gratilla in South Africa, and for the local industry to become part of an already successful and very lucrative international echinoculture industry. 

This work was supported with funding from the Woolworths/UCT Fellowship, National Research Foundation and Department of Agriculture, Forestry and Fisheries (DAFF). Special thanks go to Ms. Stacey Jordaan, Mr. Rheinhardt Scholtz and Mr. Alick Hendricks for assisting with dissections, data collection and general urchin husbandry, and to Ms.Yolanda Melo for her invaluable help with the interpretation of gonad maturity stages from histology. We would also like to thank the DAFF Aquaculture Research Facility in Sea Point, Cape Town for providing tank space and invaluable assistance, as well as Irvine & Johnson Cape Abalone farm for providing farmed Ulva. 

Agatsuma Y., Sato M. & Taniguchi K. (2005) Factors causing brown colored gonads of the sea urchin Strongylocentrotus nudus in northern Honshu, Japan. Aquaculture 249, 449–458.

Akiyama T., Unuma T. & Yamamota T. (2001) Optimum protein level in a purified diet for young red sea urchin Pseudocentrotus depressus. Fisheries Science 67, 361– 363.

Anderson R.J., Bolton J.J., Molloy F.J. & Rothman K.W.G. (2003) Commercial seaweeds in southern Africa. In: Proceedings of the 17th International Seaweed Symposium, (ed. by R.O. Chapman, R.J. Anderson, V.J. Vreeland & I.R. Davison), pp. 1–512. Oxford University Press, Oxford.

Anderson R.J., Rothman M.D., Share A. & Drummond H. (2006) Harvesting of the kelp Ecklonia maxima in South Africa affects its three obligate, red algal epiphytes. Journal of Applied Phycology 18, 343–349.

Anderson R.J., Rand A., Rothman M.D. & Bolton J.J. (2007) Mapping and quantifying the South African kelp resource. African Journal of Marine Science 29, 369–378.

AOAC International (2002) 4. Animal feed. In: Official Methods of Analysis (17th edition), (ed. by W. Horwitz), pp. 1–36. Association of Official Analytical Chemists Inc., Arlington, Virginia, USA. AOAC Official Method 992.15

Bischof K., Hanelt D. & Wiencke C. (2002) UV-radiation and Arctic marine macroalgae. In: UV Radiation and Arctic Ecosystems, (ed. by D. Hessen), pp. 227–243. Springer, New York.

Bolton J.J. (1986) Marine phytogeography of the Benguela upwelling region on the west coast of southern Africa – a temperature dependent approach. Botanica Marina 29, 251–256.

Bolton J.J., Robertson-Andersson D.V., Shuuluka D. & Kandjengo L. (2009) Growing Ulva (Chlorophyta) in integrated systems as a commercial crop for abalone feed in South Africa: a SWOT analysis. Journal of Applied Phycology 21, 575–583.

Bo¨ttger S.A., Devin M.G. & Walker C.W. (2006) Suspension of annual gametogenesis in North American green sea urchins (Strongylocentrotus droebachiensis) experiencing invariant photoperiod – applications for land-based aquaculture. Aquaculture 261, 1422–1431.

Bucke D. (1989) Histology. In: Methods for the Microbiological Examination of Fish and Shellfish, (ed. by B. Austin & D.A. Austin), pp. 69–97. Ellis Horwood, Chichester.

Byrne M. (1990) Annual reproductive cycles of the commercial sea urchin P. lividus from an exposed intertidal and a sheltered subtidal habitat on the West Coast of Ireland. Marine Biology 104, 275–289.

Cook E.J. & Kelly M.S. (2007) Enhanced production of sea urchin Paracentrotus lividus in integrated openwater cultivation with Atlantic salmon Salmo salar. Aquaculture 273, 573–585.

Cook E.J., Kelly M.S. & McKenzie J.D. (1998) Somatic and gonadal growth of the sea urchin Psammechinus miliaris (Gmelin) fed artificial salmon feed compared with a macroalgal diet. Journal of Shellfish Research 17, 1549–1555.

Dworjanyn S.A. & Pirozzi I. (2008) Induction of settlement in the sea urchin Tripneustes gratilla by macroalgae, biofilms and conspecifics: a role for bacteria? Aquaculture 274, 268–274.

Dworjanyn S.A., Pirozzi I. & Liu W. (2007) The effect of the addition of algae feeding stimulants to artificial diets for the sea urchin Tripneustes gratilla. Aquaculture 273, 624–633.

Dy D.T., Uy F.A. & Corrales C.M. (2002) Feeding, respiration, and excretion by the tropical sea urchin Tripneustes gratilla (Echinodermata: Echinoidea) from the Phillippine coral reefs. Journal of Experimental Marine Biology and Ecology 251, 227–238.

Fernandez C. & Boudouresque C.F. (1998) Evaluating artificial diets for small Paracentrotus lividus (Echinodermata: Echinoidea). In: Echinoderms: San Francisco, (ed. by R. Mooi & M. Telford), pp. 651–657. Balkema publications, Rotterdam.

Fernandez C. & Boudouresque C.F. (2000) Nutrition of the sea urchin Paracentrotus lividus (Echinodermata: Echinoidea) fed different artificial food. Marine Ecology Progress Series 204, 131–141.

George S.B. & Lawrence J.M. (2002) The effects of carotenoids on egg production and larval development of the edible sea urchin Lytechinus variegatus: the role of adult and larval diets. World Aquaculture 33, 55–61.

Halekoh U., Højsgaard S. & Yan J. (2006) The R package geepack for generalized estimating equations. Journal of Statistical Software 15, 1–11.

Hammer B.W., Hammer H.S., Watts S.A., Desmond R.A., Lawrence J.M. & Lawrence A.L. (2004) The effects of dietary protein concentration on feeding and growth of small Lytechinus variegates (Echinodermata: Echinoidea). Marine Biology 145, 1143–1157.

Hammer B.W., Hammer H.S., Watts S.A., Lawrence J.M. & Lawrence A.L. (2006) The effect of dietary protein and carbohydrate concentration on the biochemical composition and gametogenic condition of the sea urchin Lytechinus variegatus. Journal of Experimental Marine Biology and Ecology 344, 109–121.

James P.J. (2006) A comparison of roe enhancement of the sea urchin Evechinus chloroticus in sea-based and land-based cages. Aquaculture 253, 290–300.

Lawrence J.M. & Agatsuma Y. (2001) The ecology of Tripneustes. In: Edible Sea Urchins: Biology and Ecology, (ed. by J.M. Lawrence), pp. 395–413. Elsevier Science B.V., Amsterdam.

Lawrence J.M., Olave S., Otaiza R., Lawrence A.L. & Bustos E. (1997) Enhancement of gonad production in the sea urchin Loxechinus albus in Chile fed extruded feeds. Journal of the World Aquaculture Society 28, 91–96.

McBride S.C., Price R.J., Tom P.D., Lawrence J.M. & Lawrence A.L. (2004) Comparison of gonad quality factors: color, hardness and resilience, of Strongylocentrotus franciscanus between sea urchins fed prepared feed or algal diets and sea urchins harvested from the Northern California fishery. Aquaculture 233, 405–422.

Muthiga N.A. (2005) Testing the effects of seasonal and lunar periodicity on the reproduction of the edible sea urchin Tripneustes gratilla (L) in Keynan coral reef lagoons. Hydrobiologia 549, 57–64.

Olave S., Bustos E., Lawrence J.M. & Carcamo P. (2001) The effect of size and diet on gonad production by the Chilean sea urchin Loxechinus albus. Journal of the World Aquaculture Society 32, 210–214.

Pantazis P.A., Kelly M.S., Connolly J.G. & Black K.D. (2000) Effect of artificial diets on growth, lipid utilization, and gonad biochemistry in the adult sea urchin Psammechinus miliaris. Journal of Shellfish Research 19, 995–1001.

Pearce C.M., Daggett T.L. & Robinson S.M.C. (2002a) Effect of protein source ratio and protein concentration in prepared diets on gonad yield and quality of the green sea urchin, Strongylocentrotus droebachiensis. Aquaculture 214, 307–332.

Pearce C.M., Daggett T.L. & Robinson S.M.C. (2002b) Effect of binder type and concentration on prepared feed stability and gonad yield and quality of the green sea urchin, Strongylocentrotus droebachiensis. Aquaculture 205, 301–323.

Pearce C.M., Daggett T.L. & Robinson S.M.C. (2004) Effect of urchin size and diet on gonad yield and quality in the green sea urchin (Strongylocentrotus droebachiensis). Aquaculture 233, 337–367.

Phillips K., Hamid N., Silcock P., Sewell M.A., Barker M., Weaver A.M., Delahunty C. & Bremer P. (2009) Effect of manufactured diets on the yield, biochemical composition and sensory quality of Evechinus chloroticus sea urchin gonads. Aquaculture 308, 49–59.

Phillips K., Bremer P., Silcock P., Hamid N., Delahunty C., Barker M. & Kissick J. (2010) Effect of gender, diet and storage time on the physical properties and sensory quality of sea urchin (Evechinus chloroticus) gonads. Aquaculture 288, 205–215.

Plank L.R., Lawrence J.M., Lawrence A.L. & Olvera R. (2002) The effect of dietary carotenoids on gonad production and carotenoid profiles in the sea urchin Lytechinus variegatus. Journal of World Aquaculture Society 33, 127–137.

Rahman S., Tsuchiya M. & Uehara T. (2009) Effects of temperature on hatching rate, embryonic development and early larval survival of the edible sea urchin, Tripneustes gratilla. Biologia 64, 768–775.

Robertson-Andersson D., Potgieter M., Hansen J., Bolton J.J., Troell M., Anderson R., Halling C. & Probyn T. (2008) Integrated seaweed cultivation on an abalone farm in South Africa. Journal of Applied Phycology 20, 579–595.

Robinson S.M.C., Castell J.D. & Kennedy E.J. (2002) Developing suitable colour in the gonads of cultured green sea urchins (Strongylocentrotus droebachiensis). Aquaculture 206, 289–303.

Schlosser S.C., Lupatsch I., Lawrence J.M., Lawrence A.L. & Shpigel M. (2005) Protein and energy digestibility and gonad development of the European sea urchin Paracentrotus lividus (Lamarck) fed algal and prepared diets during spring and fall. Aquaculture Research 36, 972–982.

Scholtz R. (2008) Feeding preferences of Tripneustes gratilla (Linnaeus) (Echinodermata: Echinodea) and the prospect of echinoculture in South Africa. Honours project, Botany Department, University of Cape Town. UCT Bolus Library: Thesis collection (BOOK) KD RHEI.

Senaratna M., Evans L.H., Southham L. & Tsvetnenko E. (2005) Effect of different feed formulations on feed efficiency, gonad yield and gonad quality in the purple sea urchin Heliocidaris erythrogramma. Aquaculture Nutrition 11, 199–207.

Shpigel M., Marciano S., McBride S.C. & Lupatsch I. (2004) The effect of photoperiod and temperature on the reproduction of the European sea urchin Paracentrotus lividus. Aquaculture 232, 343–355.

Shpigel M., McBride S.C., Marciano S., Ron S. & Ben- Amotz A. (2005) Improving gonad colour and somatic index in the European sea urchin Paracentrotus lividus. Aquaculture 245, 101–109.

Shpigel M., Schlosser S.C., Ben-Amotz A., Lawrence A.L. & Lawrence J.M. (2006) Effects of dietary carotenoid on the gut and the gonad of the sea urchin Paracentrotus lividus. Aquaculture 261, 1269–1280.

Shuuluka D. (2011). Ecophysiological studies of three South African Ulva species from integrated seaweed/ abalone aquaculture and natural populations. PhD thesis, Botany Department, University of Cape Town.

Smith M.J. (2007) Seasonal variation in nutritional content of the kelp Ecklonia maxima on the west and south west coasts of South Africa, with reference to its use as abalone feed. M.Sc. dissertation, University of Cape Town, South Africa.

Spirlet C., Grosjean P. & Jangoux M. (1998) Closed-circuit cultivation of the edible sea-urchin Paracentrotus lividus: optimization and control of gonadal growth. In: Echinoderms. San Francisco, (ed. by R. Mooi & M. Telford), pp. 835. Balkema, Rotterdam.

Spirlet C., Grosjean P. & Jangoux M. (2000) Optimization of gonad growth by manipulation of temperature and photoperiod in cultivated sea urchins, Paracentrotus liI´idus (Lamarck) (Echinodermata). Aquaculture 185, 85–99.

Stegenga H., Bolton J.J. & Anderson R.J. (1997) Seaweeds of the South African West Coast. In: Contributions from the Bolus Herbarium, (ed. by A.V. Hall), Vol. 18, pp. 1–655. Bolus Herbarium, University of Cape Town, Cape Town.

Stephenson T.A. (1948) The constitution of the intertidal flora and fauna of South Africa III. In: Annals of the Natal Museum, Vol. 11 pp. 207–324. Natal Museum, Pietermaritzburg.

Symonds R.C., Caris-Veyrat C., Kelly M.S. & Young A.J. (2007) Carotenoids in the sea urchin Paracentrotus lividus: occurrence of 9′-cis echinenone as the dominant carotenoid in gonad colour determination. Comparative Biochemical and Physiology, Part B Biochemical and Molecular Biology 148, 432–444.

Taylor A.M., Powell M.L., Watts S.A. & Lawrence A.L. (2009) Formulated feed supports weight gain and survivorship in juvenile sea urchins Lytechinus variegatus. Journal of the World Aquaculture Society 40, 780–787.

Troell M., Robertson-Andersson D.V., Anderson R.J., Bolton J.J., Maneveldt G., Halling C. & Probyn T. (2006) Abalone farming in South Africa: perspectives on kelp resources, abalone feed, potential for on-farm seaweed production and socio-economic benefits. Aquaculture 257, 266–281.

Vaitilingon D., Rasolofonirina R. & Jangoux M. (2003) Feeding preferences, seasonal gut repletion indices, and diel feeding patterns of the sea urchin Tripneustes gratilla (Echinodermata: Echinoidea) on a coastal habitat off Toliara (Madagascar). Marine Biology 143, 451–458.

Walker C.W. & Lesser M.P. (1998) Manipulation of food and photoperiod promotes out of season gametogenesis in the green sea urchin, Strongylocentrotus droebachiensis: implications for aquaculture. Marine Biology 132, 663–676.

Woods C.M.C., James P.J., Moss G.A., Wright J. & Siikavuopio S. (2008) A comparison of the effect of urchin size and diet on gonad yield and quality in the sea urchin Evechinus chloroticus Valenciennes. Aquaculture International 16, 49–68.