CERMES Technical Report No. 87

The effectiveness of a modified turtle excluder device (TED) in reducing the bycatch of elasmobranchs in the Atlantic seabob (Xiphopenaeus kroyeri) industrial trawl fishery of Guyana

A. GARSTIN, H.A. OXENFORD AND D. MAISON

Centre for Resource Management and Environmental Studies (CERMES) Faculty of Science and Technology, The University of the West Indies Cave Hill Campus, Barbados 2017

ABSTRACT

The Atlantic seabob (Xiphopenaeus kroyeri) trawl fishery is extremely important to Guyana, with some 88 licensed industrial trawling vessels harvesting around 15,000 mt per year, almost all of which is exported to the US and EU, representing Guyana’s most valuable seafood export. The key player in this industry, the Guyana Association of Private Trawler Owners and Seafood Processors (GAPTO&SP) is taking pro-active steps in pursuing Marine Stewardship Council certification for the seabob trawl fishery to ensure top market prices and long-term sustainability of the seabob stock. To this end, all commercial vessels in the fleet are using turtle excluder devices (TEDs) and bycatch reduction devices (BRDs) in their trawl nets. However, the effectiveness of these devices in reducing the bycatch of vulnerable sharks and rays has not yet been examined. This study, requested by GAPTO&SP, represents the first attempt to document the bycatch of these discarded species by the seabob trawl fleet, and to compare the effectiveness of two different TED designs. Over the period July-August 2014, five trips were taken on three different seabob vessels to document the species, sizes and condition of all sharks and rays landed and discarded during the normal 24 hour-day operation of the vessels. Wherever possible, vessels deployed nets fitted with a standard control TED, simultaneously with nets fitted with a modified test TED for comparison of bycatch by the two gears. A total of 131 tows were sampled, 80 of which represented simultaneous tows of both the control and modified test TEDs. Shark and ray bycatch in the nets with control TEDs comprised eight ray species and one shark species including, among the infrequently landed species, three ‘Near Threatened’ and one ‘Critically Endangered’ ray species according to the IUCN Red List. The use of the modified TEDs significantly reduced the overall mean size of individuals in the elasmobranch bycatch by 6.3%. Most importantly the mean sizes of the two ray species representing >80% by number of elasmobranchs taken as bycatch were reduced by 9.4% and this resulted in a near elimination of mature females in the bycatch. By excluding larger individuals from the cod-end, a statistically significant and substantial decline in the catch rate (by 40%) was observed when using the test TED compared with the control TED (mean elasmobranch bycatch rate dropped from 2.29 to just 1.37 individuals per twin-net hr-1). This also resulted in a virtual elimination of the three ‘Near Threatened’ ray species in the bycatch, although it had little effect on the capture of the small-sized ray and shark species, including the ‘Critically Endangered’ Bancroft’s numbfish. We conclude that the modified TED was very successful in reducing important elements of the elasmobranch bycatch and should advance the progress towards attaining the bycatch standards required for MSC certification. We further recommend that the impact of this TED on the capture rate of the target seabob, which was beyond the scope of this study, should be examined to inform the decision on mandatory adoption of this gear modification.

Keywords: elasmobranch bycatch; Guyana, MSC certification, seabob, industrial trawl fishery

i LIST OF ACRONYMS

BRD Bycatch reduction device

CRFM Caribbean Regional Fisheries Mechanism

FAO Food and Agriculture Organization

GAPTO&SP Guyana Association of Private Trawler Owners and Seafood Processors

IUU Illegal, unreported and unregulated

MSC Marine Stewardship Council

TED Turtle excluder device

VMS Vessel monitoring surveillance

WECAFC Western Central Atlantic Fishery Commission

ii ACKNOWLEDGEMENTS

We gratefully acknowledge the special contributions of Captain Naresha Ramanand and the crew of FV Pacuma, Captain Errol Adams and the crew of FV Ranmar 6, and Captain Nerwanie Persaud and the crew of FV Rebel’95 for allowing AG to join the crew on commercial fishing trips and agreeing to fish with both TED designs simultaneously. The research would not have been possible without this generous collaboration. We also acknowledge the generous support and assistance of the Guyana Association of Private Trawler Owners and Seafood Processors (GAPTO&SP), the management and staff of Noble House Seafoods, and the management and staff of Pritipaul Singh Investment Inc.

We further acknowledge CERMES for their funding support through a research grant to AG, and the Ministry of Agriculture, Government of Guyana for their assistance and approval of the research. We also thank Patrick McConney of CERMES, Tomas Willems of the Institute for Agricultural and Fisheries Research, Suriname, and Mark Nijhof of Heiploeg International, Netherlands for their interest and constructive critique of the study.

iii Table of Contents 1 INTRODUCTION ...... 1 1.1 Commercial importance of seabob ...... 1 1.1.1 Guyana’s seabob fishery ...... 1 1.2 Management of the seabob resources...... 2 1.3 Importance of the bycatch issue ...... 2 1.3.1 International policy on bycatch ...... 3 1.3.2 Vulnerability of the elasmobranchs ...... 3 1.4 Market standards for shrimp ...... 3 1.4.1 Marine Stewardship Council certification ...... 4 1.4.2 Guyana Association of Private Trawler Owners and Seafood Processors ...... 4 1.4.3 Progress towards certification ...... 4 1.5 Study aims and objectives ...... 5 2 METHODS ...... 5 2.1 Survey area ...... 5 2.2 Vessels, gear and fishing operations ...... 6 2.3 At-sea data collection ...... 8 2.3.1 Physical data ...... 8 2.3.2 Biological data ...... 8 2.4 Data handling and analysis ...... 9 3 RESULTS ...... 11 3.1 Sampling effort ...... 11 3.2 Shark and ray bycatch using standard control TEDs ...... 11 3.2.1 Species composition and individual size ...... 11 3.2.2 Gender and life condition...... 15 3.2.3 Catch rates ...... 21 3.3 Shark and ray bycatch using modified test TEDs ...... 22 3.3.1 Species composition and individual size ...... 22 3.3.2 Gender and life condition...... 23 3.3.3 Catch rates ...... 23 3.4 Comparison of shark and ray bycatch between control and test TEDs ...... 24 3.4.1 Species composition and individual size ...... 24 3.4.2 Gender and life condition...... 26 iv 3.4.3 Catch rates ...... 27 4 DISCUSSION ...... 28 5 Conclusion ...... 32 6 REFERENCES ...... 34

Cover photographs: A. Garstin

Citation: Garstin, A., H.A. Oxenford and D. Maison. 2017. The effectiveness of a modified turtle excluder device (TED) in reducing the bycatch of elasmobranchs in the Atlantic seabob (Xiphopenaeus kroyeri) industrial trawl fishery of Guyana. Centre for Resource Management and Environmental Studies, The University of the West Indies, Cave Hill campus, Barbados. CERMES Technical Report No. 87: 36pp.

v 1 INTRODUCTION

1.1 Commercial importance of seabob

The Atlantic seabob (Xiphopenaeus kroyeri) is a penaeid shrimp species found only in the western Atlantic; its range extending from the southern United States, across the Gulf of Mexico and the Caribbean coasts of Central America, to the continental shelf of South America as far south as Brazil (the Guianas-Brazil shelf) (Taveres 2002). Seabob supports artisanal and industrial trawl fisheries across this range, with large-scale industrial fisheries established in Venezuela, the Guianas and Brazil (Holthuis 1980), although trawl fisheries in Venezuela have been shut down since 2009 (Earl 2009). The seabob trawl fishery is one of the most productive fisheries for Guyana, Suriname and French Guiana (WECAFC 2013). The global reported capture of seabob reached 52,000 MT in 2012, more than half of which (24,883 MT), was from Guyana (FAO FishStat global capture production database http://data.fao.org ).

1.1.1 Guyana’s seabob fishery

The marine fisheries of Guyana comprise an offshore industrial (trawl) fishery of registered vessels (numbering around 100), a semi-industrial deep slope fishery for red snapper (numbering less than 50 vessels) and an inshore artisanal fishery with only partial registration of the fleet (numbering around 1,200 vessels). The fisheries have been described in various research papers and national reports (e.g. Shepherd et al. 1999, FAO 2005, Greer 2005, Maison 2007, 2016 Richardson 2013, MacDonald et al 2015, Kalicharan in prep.) and information pertaining to the seabob fishery is summarised here from these descriptions.

Seabob are targeted by both the inshore artisanal and the offshore industrial trawl fleets, although the former are estimated to account for just 2% of Guyana’s total seabob landings, whilst the remaining 98% are taken by the industrial fleet. The only vessels used in the inshore shrimp fishery are the Chinese seine vessels and they target seabob together with white belly shrimp and various finfishes. The vessels are relatively small (6-12 m in length), flat-bottomed dory-type wooden boats powered by sail or outboard engine and are believed to number around 250-300 vessels (307 Chinese seine vessels in 2011) and represent just over a quarter (27%) of the inshore artisanal fleet. They use funnel-shaped fyke nets attached to poles and set them in or near the river mouths to utilise the strong currents. They generally set one net per trip, but may haul multiple times during a trip. Each trip lasts approximately 6-12 hours and will fish an average of 18 trips a month (200-300 trips a year).

Offshore industrial trawling for seabob began in Guyana in the mid-1980s. This fleet is reserved for Guyanese fishers and there are currently 88 locally-owned, licenced seabob trawlers. The vessels are steel-hulled, standard Gulf of Mexico type trawlers of 19-23 m in length and powered by inboard diesel engines. They mainly target seabob and some finfish species: bangamary (Macrodon ancylodon) and butterfish (Nebris microps), with small quantities of other penaeid shrimps taken as by-catch. Since the late 1990s they have been using a twin rig set-up with a sledge to tow four nets simultaneously. The nets are approximately 11-15 m in length, have a mesh size of 4-5 cm in the wings and 2.5-3.5 cm in the cod-end and are fitted with mandatory turtle extraction devices (TEDs) and drop chains around the mouth to improve catch quantity. 1 They also use a small try-net (3 m long) for short periods of time to test the area for abundance of shrimp. The fleet generally operates 15-30 km from shore in 18-20 m of water depth on muddy substrates and moves to different areas during the fishing season. There is a 6-week closed season sometime between August and October (decided each year according to the drop- off in catch rate). The trawlers make on average about 2-3 trips a month (30 per year). Trips vary in length from 3-4 days at the peak of the season (~Dec-Feb) to 8-10 days when fishing is poorer. They fish 24 hr a day generally making 4-6 hauls per day of 3-4 hr duration and operate with a crew of six. The seabob are landed fresh on ice. Between 2010-2012, 45% of the total seabob landings were exported, mainly to US markets (averaging 12,509 mt per year and representing a value of US$30.1 million per year), accounting for around 0.8-1% of the country’s GDP over the same period (Richardson 2013). The most recent published estimates suggest that almost all of the annual harvest of around 15,000 mt is exported to the US and EU as frozen shell-off tails (product weight around 5,000 mt) and is valued at US$45 million per year, representing Guyana’s most valuable seafood export (Maison 2016).

1.2 Management of the seabob resources

The shared shrimp and finfish resources of the Guianas-Brazil shelf have received considerable attention over the last few decades through the work of various joint scientific working groups supported by the Food and Agriculture Organization-Western Central Atlantic Fishery Commission (FAO-WECAFC) and Caribbean Regional Fisheries Mechanism (CRFM). This includes the FAO-WECAFC Ad Hoc Shrimp and Groundfish Working Group of the Guianas- Brazil Shelf (initiated in 1986) and in partnership with the CRFM Shrimp and Groundfish Scientific Working Group (now CRFM Continental Shelf Scientific Working Group) since 1996. These working groups have undertaken a number of stock assessments, with the most recent stock assessment being conducted in 2012/2013 (CRFM 2013).

Stock assessments in the early 2000s raised concerns about the sustainability of the industrial trawl fishery operating at that time, and indicated a need for better management, including a reduction in fishing effort (CRFM 2007). In response, the industrial seabob fleet size in Guyana has been reduced by approximately 20% and Guyana is currently developing an improved fishery management plan for the period 2013-2017 under an EU-ACP Fish II Project. This will include a requirement for all seabob vessels to be equipped with vessel monitoring surveillance (VMS) systems and additional by-catch reduction devices (BRDs) in line with the requirements for Marine Stewardship Council (MSC) certification (SOFRECO 2013; Richardson 2013). The most recent stock assessment for seabob was completed in 2013 and the primary stakeholders are in the final stages of developing harvest control rules for this fishery (Medley 2014).

1.3 Importance of the bycatch issue

Bycatch is defined by Hall (1996) as “all non-target fish whether retained and sold or discarded.” Shrimp trawl fisheries are well known to have one of the highest bycatch rates of any fishery and have been the subject of numerous studies and technical innovations to reduce bycatch (Earys 2007). The FAO maintains a position that bycatch “threatens the long-term sustainability of many fisheries and the maintenance of biodiversity in many areas, resulting in increased food

2 insecurity and adversely affecting the livelihoods of millions of fishers and fishworkers dependent on fish resources” (FAO 2011).

1.3.1 International policy on bycatch

Global leaders have recognised the importance of bycatch as a management issue to be urgently addressed and several initiatives put forth by the FAO seek to address their concerns. The FAO Code of Conduct for Responsible Fisheries, adopted in 1995, promotes the sustainable use of the marine environment by minimising fisheries impacts on non-target species. The move towards an ecosystem approach to fisheries (EAF) and the guidelines on EAF (FAO 2003) also highlight the importance of impacts to other members and parts of the ecosystem. In 1999, the FAO established the International Plan of Action for the Conservation and Management of Sharks (IPOA-Sharks) which highlighted the special vulnerability of this group to fishing and as a bycatch species. More recently, in 2011, the International Guidelines on Bycatch Management and Reduction of Discards, specifically outlines policies that should be established by countries to further reduce their fisheries’ impacts on non-target species. One specific aspect of these guidelines is a call for fisheries to “minimize bycatch through the modification of fishing gears and practices” (FAO 2011). The current study represents a small step in this direction by investigating the effectiveness of using a modified TED in the Guyanese industrial seabob trawl fishery to further reduce the catch of non-target species, specifically the elasmobranchs (sharks and rays).

1.3.2 Vulnerability of the elasmobranchs

Of particular concern is the bycatch of elasmobranchs (sharks and rays) because of their k- selected life-histories, which makes them particularly susceptible to rapid declines in stock size (Stevens et al. 2000). This is because k-selected species are slow-growing, reach maturity at a late age, have low fecundity and are long-lived (Camhi et al 1998). The International Union for Conservation of Nature (IUCN) has recognised this threat, most recently by reporting that 25% of sharks and rays are threatened with extinction, with rays being the more susceptible of the two (IUCN 2014). The main cause in the decline of shark and ray populations has been linked to fishing pressures; for most of the world sharks and rays are captured unintended. With their flattened bodies and larger size, sharks and rays are expected to escape through the turtle excluder device (TED), a bycatch reduction grid that forces large marine organisms out of the trawl net (Stobutzki et al 2002; Brewer et al 2006; Sala, Lucchetti and Affronte 2011).

1.4 Market standards for shrimp

Seabob, being primarily an export fishery product, must meet international market requirements for successful export. For example, as of 1989, the US market requires all shrimp fisheries that export to the United States to utilise TEDs in their nets (NMFS 2014). The European market, which generally pays a higher price, has more stringent demands including product traceability, especially with the passage of the 2010 Illegal, Unreported and Unregulated (IUU) legislation. Its intent is to remove seafood products from unsustainable fisheries from the market, by restricting countries that do not adhere to reporting procedures, punishing vessels that engage in illegal fishing activities and helping developing countries legislate national fisheries policies 3 (European Commission 2014). Furthermore, other niche markets have emerged, over the last decade or so, that pay premium prices for ‘eco-friendly’ or ‘sustainable seafood’ products. One of the leaders in developing a globally accepted eco-label is the Marine Stewardship Council (MSC), which began as an initiative in 1996 by the World Wildlife Fund and Unilever, as a response to the Grand Banks cod fishery collapse of 1992 (WWF 2011).

1.4.1 Marine Stewardship Council certification

MSC certification seeks to ensure that marketed seafood comes from sustainably managed fisheries. The organisation addresses the sustainability of fishing pressure on the target species, minimising the environmental impact from fishing practices and providing effective management of the fishery (MSC 2014). A key issue of concern for sustainable management, typical of any trawl fishery, but particularly regarding the MSC certification which examines ecosystem-level impacts of fishing, is that of bycatch reduction. As part of the MSC assessment process, several considerations are specifically made for bycatch: “MSC fisheries will no longer be at risk of generating cumulative negative impacts on bycatch species” and “fisheries will need to regularly review alternative measures that could reduce the mortality of unwanted species in the catches” (MSC 2014). The research undertaken by this study should therefore help address the fishery’s need to review novel approaches to mitigating bycatch during seabob trawling, thereby enhancing their application for MSC certification. MSC certification would afford Guyana preferential access to more and higher priced markets in the USA and Europe (MSC 2014).

1.4.2 Guyana Association of Private Trawler Owners and Seafood Processors

The Guyana Association of Private Trawler Owners and Seafood Processors (GAPTO&SP) was established by the industry and includes the major entities from the industrial trawl fishery: Noble House Seafoods, Pritipaul Singh Investment Inc, Guyana Quality Seafoods, BEV Processors Inc. and several independent trawler owners (CRFM 2013). The GAPTO&SP has a significant interest in maximising profits in the seabob fishery, whilst recognising the importance of a sustainable resource, and is therefore currently pushing for reforms that would allow MSC certification for their fishery.

1.4.3 Progress towards certification

Both the Guyana and Suriname seabob fisheries underwent a pre-assessment to the MSC standard in 2008/9 and thereafter initiated fishery improvement programmes to address identified problems (Maison 2016). Suriname went under full assessment in 2009 and subsequently received MSC certification of their industrial seabob trawl fishery in 2011 (FCI 2011, Maison 2016).

Some of the main differences between the Guyana and Suriname seabob fisheries are the fact that the Guyanese fishery involves a much larger number of fishery stakeholders, has a much bigger fleet, more owners and a larger number of processors than the Surinamese fishery. As such, addressing the problems outlined in the MSC pre-assessment is a much greater task. However, GAPTO&SP with the support of the Guyanese government, has signed up to, and has continued these efforts with a formal agreement establishing the Guyana Seabob Trawl Fishery

4 Improvement Plan in May 2012. By August the Guyana Seabob Working Group was formed under the chairmanship of the Department of Fisheries, with the support of a range of other government departments and GAPTO&SP (Maison 2016).

Under a much improved management plan for 2013-2017, all industrial Guyanese seabob trawling vessels are now required to carry vessel monitoring surveillance (VMS) tracking devices and to install bycatch reduction devices (BRDs) in the trawl nets in addition to the TEDs previously mandated for export to the USA (SOFRECO 2013; Richardson 2013). However, a recent study of the ray bycatch in the neighbouring Surinamese seabob fishery, where the rigging and TED design is similar to the gear used in Guyana, indicated that although the use of TEDs and BRDs has reduced the ray bycatch by 36% (Willems et al 2013) rays, especially smaller- sized individuals, are still being caught in every haul at a mean rate of 15.3 individuals hr-1. This information has prompted the GAPTO&SP to investigate their own shark and ray bycatch rates and to experiment with a modified TED design. The University of the West Indies was therefore approached by GAPTO&SP to assist them with conducting the current study.

1.5 Study aims and objectives

As there is currently no reporting on the bycatch discards by the Guyanese industrial seabob trawl fishery, the aim of this research is to assist the GAPTO&SP in assessing the efficacy of a modified TED design in reducing the bycatch of sharks and rays. This was carried out by:

 Determining the species composition, size frequency, gender and condition of all sharks and rays taken as bycatch during standard fishing operations (with control TEDs)

 Determining the catch rates of all species of sharks and rays taken during standard fishing operations (with control TEDs)

 Comparing the catch composition and catch rates of all sharks and rays taken during standard fishing operations with control TEDs versus modified TEDs

2 METHODS

This study was conducted over the summer of 2014 (July-August) through on-board measurements and observations of shark and ray bycatch taken during the typical industrial seabob fishing operations, using double twin rigs with standard trawl nets fitted with the standard TED, fished simultaneously alongside standard nets carrying modified TEDs.

2.1 Survey area

Fishing trips were conducted on the Guyanese continental shelf, in the area normally utilised by the industrial seabob fishing fleet (Figure 1, see also Maison 2016). Choice of trawling locations was at the discretion of the fishing vessel captains and should therefore represent the typical catch found during normal fishing practices. The bottom environment in this area is characterised by a depth between 14 to 30 m with a substrate consisting mainly of mud (SOFRECO 2013).

5 2.2 Vessels, gear and fishing operations

The study was conducted on typical GAPTO&SP commercial outrigger trawlers, fitted for double twin-rig bottom trawling. These vessels are approximately 21 m in length and have a 450- hp engine capacity (Figure 2). Double twin-rig bottom trawling involves towing a mid-trawl sledge on each side of the vessel to keep the nets open and separated. Two steel-framed wooden doors are connected to each sledge, which are in turn fitted to two distinct net pairs. The nets used by the fleet have a 4-5 cm mesh size in the wings, which is reduced down to 2.5 cm in the cod-ends.

The Guyanese industrial seabob trawlers are now required by the Ministry of Agriculture to use both BRDs and TEDs in each of the four cod-ends. The BRDs used in the Guyanese seabob trawl nets comprise a large mesh panel (10 x 10 meshes, 10 cm mesh size) placed on the upper surface of the net, near the cod-end (Figure 3) which is designed to allow finfish and other unwanted bycatch species to exit the trawl net, whilst retaining shrimps (Crespi and Pardo 2012).

Essequibo River

Figure 1. Satellite image of the coastal area of Guyana showing the approximate area used by the industrial fleet. The star signifies the landing site for the fleet. Inset shows location of Guyana on the north coast of South America.

6 (a) (b)

Figure 2. (a) Typical double twin-rig bottom trawler of the Guyanese industrial seabob fleet and (b) diagram of double twin-rig bottom trawler to illustrate the arrangement of the four trawl nets fished simultaneously. Note that a much smaller ‘try net’ (not shown) is also towed off the centre of the stern to inform the captain of the catch rate. Diagram source: Broadhurst (2004).

TED BRD panel Cod-end

Escape flap

Figure 3. Diagram of cod-end fitted with slanted, rigid TED forward of the BRD mesh panel. Image adapted from Willems et al (2013).

The TED is a more specialized bycatch reduction device for trawl nets, specifically designed to allow turtles and other large species to exit the net with minimal harm (Brewer et al 2006; Cox et al 2007; Crespi and Prado 2012). For most TEDs, a large metal grid is fitted across the trawl net at an angle. This allows small organisms to pass through the grid and on into the net cod-end, but re-directs larger organisms that cannot pass through the grid to an escape panel or flap (Figure 3). In the Guyanese seabob trawl fishery, the aluminium TEDs are 86.4 cm x 106.7 cm (34” x 42”) and constructed with 1.3 cm (0.5”) thick bars. They are required to have a maximum bar spacing of 10.2 cm (4”). Whilst some vessels use the maximum spacing, others use smaller spacing of 8.9 cm (3.5”).

For a comparative control, a mixture of 8.89 cm (3.5”) and 10.16 cm (4”) bar spacing TEDs were utilised during sampling. When available, the test TEDs that were used had a considerably reduced bar spacing of 4.45 cm (1.75”) as well as a horizontal brace bar (Figure 4).

7

Figure 4. Photograph of the modified test TEDs used in this study, with a bar spacing of 4.45 cm (1.75”) and a horizontal brace bar.

Trawlers operate on a 24-hour schedule at sea. Trawls are conducted with simultaneous deployment of four cod-ends (Figure 2b) and a try-net, which allows the captain to approximate the size and composition of the catch in the cod-ends. This try-net is brought up hourly to monitor the progress of the trawl. A typical trawl will last four hours, at which point the four cod-ends are brought to the surface and hoisted on deck. However, the duration varies throughout the season according to the availability of seabob. The nets are redeployed while the crew begins the sorting process, which typically lasts between one to four hours.

2.3 At-sea data collection

At-sea data collection was carried out on several industrial vessels between July and August 2014. Trawls were monitored on a continuous, 24-hour basis.

2.3.1 Physical data

At the start and conclusion of each trawl, the GPS coordinates and time were documented, as well as the towing speed for the duration of the trawl, using a handheld GPS receiver (Garmin GPS 72H). Weather conditions were also recorded for each trawl at approximately midway through the trawl.

2.3.2 Biological data

Control TEDs were fitted to the two port trawl net cod-ends and test TEDs (whenever available) were fitted to the two starboard cod-ends. At the end of each haul, all nets where brought up and emptied on deck, assuring separation between catches of the starboard test-nets and the port

8 control-nets, but it was not possible to separate the catch of the two nets on each side. As such, a double twin rig trawl provided two sample replicates; one from the twin nets on the port side and the other from the twin nets on the starboard side.

Sharks and rays were manually sorted from the catch. All sharks and rays were identified to species level using the FAO species guides for Area 31 (Compagno 2002; McEachran and de Carvalho 2002). All ray individuals were measured with a flexible tape measure (maximum disc width) to the nearest centimetre; all shark individuals were measured with a flexible tape measure (fork length and girth) to the nearest centimetre (Figure 5). The sex of each individual was then noted by observing the absence or presence of claspers, the latter signifying a male specimen. Life condition (good, poor or dead) was also noted (subjectively critiqued) and the ray or shark was then released if found alive.

girth

disc width fork length

clasper

Figure 5. Size measurements taken for shark (left) and ray (right) individuals taken as bycatch in the seabob trawls. The difference in morphology between a female (A) and a male (B) ray is shown. Images adapted from: Florida Museum and Natural History at www.flmnh.ufl.edu/fish/education.

2.4 Data handling and analysis

Data collected at sea were initially recorded on water-resistant paper. On return to shore, all data were transferred to an Excel database for further descriptive and statistical analyses. ArcGIS 10.1 was utilised to generate visual representations of spatial data.

For general descriptions of bycatch rate, species and size composition, gender and condition, all trawl data were used. For comparison of catch rates between the two TED designs, only data from simultaneous trawls were used (i.e. those when control and test TEDs were fished simultaneously). Furthermore, when comparing individual species catch rates, sizes, gender and condition between the two TED designs, only species that accounted for at least three percent of shark and ray bycatch were considered, to avoid problems associated with small sample sizes.

To analyse differences in mean size of all sharks and rays taken by control TEDs versus test TEDs, two sample t-tests were used for the three most numerous species: smooth butterfly ray, longnose stingray and smalleye smooth-hound, and the total ray disc width. The sample sizes for these species were large enough in both the control and test TED to allow for testing.

9 To assess differences in sex ratios of bycatch between the two TED designs, chi square contingency tests were utilised for the three most common bycatch species caught: longnose stingray, smooth butterfly ray and smalleye smooth-hound. Similarly, a chi square contingency test was also used to evaluate the differences in life condition percentages between the two TED designs. As the total number of live sharks and rays caught was small, they were pooled for this analysis.

Shark and ray bycatch rates per trawl were calculated as the number of individuals caught per twin net haul per hour. Parametric statistical testing was used wherever assumptions of normal distribution and homogeneity of variance were satisfied. If not, then equivalent non-parametric procedures were employed. Differences in mean bycatch rates between the control and test TEDs that were fished simultaneously were assessed by paired t-tests using individual trawl data. The effect of TED type on sex ratio was tested with Pearson’s chi-square contingency tests for independence.

10 3 RESULTS

3.1 Sampling effort

A total of five fishing trips were monitored between July 12 and August 29, 2014 on three different vessels belonging to GATPO&SP members: FV Pacuma, FV Rebel’95 and FV Ranmar 6 (Table 1). This added up to a total of 31 days of 24 hr at-sea sampling, and represented 528 hours and 20 minutes of trawling time or a total of 1,056 hours and 40 minutes of twin-net trawling hours.

Table 1. Summary of fishing trip data.

Mean towing Mean length of No. trawl replicates Dates No. of trawls speed trawl (twin-net hauls) Days (double twin-rig) Control Test Start End knots SE hr:min min SE TED TED 12-Jul-14 19-Jul-14 8 41 2.4 0.05 3:42 222 3.5 82 0 24-Jul-14 1-Aug-14 9 47 2.2 0.02 3:50 230 4.1 47 47 10-Aug-14 15-Aug-14 6 20 3.3 0.06 4:46 286 5.6 30 10 18-Aug-14 22-Aug-14 5 17 3.1 0.03 4:14 254 17.1 17 17 27-Aug-14 29-Aug-14 3 6 2.3 0.11 4:49 289 16.8 6 6 Overall 31 131 2.6 0.04 4:02 242 3.7 182 80

A total of 131 double twin-rig trawls with an average duration of 4 hours 2 minutes and a mean towing speed of 2.6 knots were sampled, comprising 262 twin-net replicate samples (Table 1). Variation in lengths and speeds of trawls was negligible across all trips as indicated by low standard error (SE) values in all cases (Table 1). A total of 51 trawls (102 twin-net replicates) were conducted using only control TEDs in all of the nets. A further 80 simultaneous catch- comparison trawls (160 replicates) where completed in which control TEDs were fitted in both cod-ends of the twin-nets on the port side (control-nets) and the twin starboard nets were fitted with test TEDs (test-nets) (Table 1). This accounts for 182 replicate trawls using the control TED and 80 replicate trawls using the test TED. Trawlers remained within the boundaries of 6.33394°N to 7.73781°N and 57.11912°W to 58.49553°W during trials, indicating that the fishing grounds ranged between 6.5 and 27 nm (12-50 km) from the Guyanese coastline (Figure 6).

3.2 Shark and ray bycatch using standard control TEDs

3.2.1 Species composition and individual size

All sharks and rays

A total of 1,399 elasmobranchs (108 sharks, 1,291 rays) were caught across four orders, seven families and nine species. These included eight ray species: Bancroft’s numbfish (Narcine bancroftii), chola guitarfish (Rhinobatos percellens), cownose ray (Rhinoptera bonasus),

11

Figure 6. Satellite image, repeated from Figure 1, showing the start and end coordinates (as white, filled circles) for all monitored seabob trawls during the study in July-August 2014. Inset shows location of Guyana on the north coast of South America. longnose stingray (Dasyatis guttata), sharpsnout stingray (Dasyatis geijskesi), southern stingray (Dasyatis americana), smooth butterfly ray (Gymnura micrura) and smalleyed round stingray (Urotrygon microphthalmum); and one shark species: smalleye smooth-hound shark (Mustelus higmani) (Table 2, Figure 7).

Table 2. List of shark and ray (elasmobranchii) bycatch species taken by the Guyanese seabob trawl fishery using standard control TEDs (n=182 twin-net hauls) and test TEDs (n = 80 twin-net hauls).

IUCN Control Test Group Order Family Species Common name n category* TED TED Torpediniformes Narcinidae Narcine bancroftii Bancroft's numbfish CR 15   Rajiformes Rhinobatidae Rhinobatos percellens Chola guitarfish NT 6   Myliobatidae Rhinoptera bonasus Cownose ray NT 8   Smalleyed round Urotrygonidae Urotrygon microphthalmum LC 56   Rays stingray Myliobatiformes Gymnuridae Gymnura micrura Smooth butterfly ray DD 1187   Dasyatis geijskesi Sharpsnout stingray NT 51   Dasyatidae Dasyatis guttata Longnose stingray DD 366   Dasyatis americana Southern stingray DD 5   Smalleye smooth- Sharks Carcharhiniformes Triakidae Mustelus higmani LC 140   hound shark * CR - critically endangered; NT – near threatened; LC - least concern; DD - data deficient

12

10 cm 15 cm 10 cm

(a) Bancroft’s numbfish (b) Chola guitarfish (c) Cownose ray

30 cm (d) Longnose stingray (e) Sharpsnout stingray (f) Southern stingray

10 cm 5 cm 10 cm

(g) Smooth butterfly ray (h) Smalleyed round stingray (i) Smalleye smooth-hound shark

Figure 7. Shark and ray species found as bycatch in the Guyanese industrial seabob trawl fishery. (a) – (h) show the eight ray species, (i) shows the single shark species. Images (d) and (f) were sourced from McEachran and de Carvalho (2002).

Most of the rays captured were small; the mean disc width for rays that passed through the control TED being 24.48 cm (± 0.26 cm), with the largest proportion (24%) between 19-21 cm (Figure 8a, Table 3). The smallest ray captured had a disc width of 4 cm and the largest ray had a disc width of 79 cm. Thus the vast majority of specimens are considerably wider than the bar- spacing of between 8.9-10.2 cm, and must therefore be passing through sideways. The only shark species is shown in the next section.

13 (a) (b)

Figure 8. Size frequency distribution of all ray species (shown by 3 cm disc width size groups) taken as bycatch in nets fitted with (a) control TEDs and (b) test TEDs.

Commonly caught species

One shark and four ray species together accounted for 98% of the elasmobranch bycatch when using control TEDs. The smooth butterfly ray alone accounted for more than half (61%) and the longnose stingray for almost a quarter (22%) (Figure 9a, Table 3). The other three common bycatch species each accounted for at least three percent of the total elasmobranch bycatch: smalleye smooth-hound shark (8%), smalleyed round stingray (3%) and sharpsnout stingray (3%). The remaining 2% of the elasmobranch bycatch comprised four species.

Figure 9. Relative species composition (% by number) of shark and ray bycatch taken in twin-net hauls with (a) control TEDS and (b) test TEDs. NB. Sharpsnout stingray is listed as ‘Near Threatened’ on the IUCN redlist.

14 Of the four ray species commonly found in the bycatch, three had a mean disc width greater than 10 cm: smooth butterfly ray (26.6 cm), longnose stingray (20.0 cm) and sharpsnout stingray (32.5 cm) (Table 3). The fourth species, smalleyed round stingray, had a mean disc width that was very close to the bar spacing of the control TEDs (9.4 cm). The sharks (smalleye smooth- hound) were also very small; the mean fork length (FL) of individuals caught in the control net being 22.3 cm and the girth averaging 8.2 cm (Table 3).

Table 3. Summary of mean sizes of shark and ray species taken as bycatch, shown separately for nets with control TEDs and test TEDs.

TED design Control Test Mean width Mean width Species SE n SE n (cm) (cm) Longnose stingray 20.02 0.33 313 18.13 0.31 53 Sharpsnout stingray1 32.50 2.12 44 28.43 1.63 7

Smalleyed round stingray 9.36 0.43 45 11.45 1.25 11

Rays Smooth butterfly ray 26.55 0.29 860 24.20 0.27 327 Total ray species 24.48 0.26 1291 22.95 0.28 403 Mean FL Mean FL SE n SE n (cm) (cm)

Smalleye smooth-hound 22.30 0.41 108 21.84 0.63 32 shark Mean girth Mean girth Sharks SE n SE n (cm) (cm) 8.21 0.15 108 7.91 0.29 32 1 indicates IUCN Near Threatened species.

The size frequency distributions for the three most commonly caught elasmobranch bycatch species are shown in Figure 10. Longnose stingray were found between 13 and 58 cm disc width, with the largest proportion in the 16-18 cm size class (44%) (Figure 10a). As females for this species are reported to mature at a disc width of 50-55cm, it is likely that just 0.5% of the female bycatch may represent mature individuals. For smooth butterfly ray, the size range was between 16 and 64 cm disc width; the most numerous size class for this species being 22-24 cm (28%) (Figure 10c). As females of this species are reported to mature at a much smaller disc width of 34-36 cm (Yokota and Lessa 2007), it appears that just around 10.9% of the female smooth butterfly rays taken as bycatch in the nets fitted with control TEDs are of reproductive age, whilst the rest are immature. Smalleye smooth-hound shark were found to range between 16 to 47 cm FL, with 56% in the 19-21 cm size class (Figure 10e).

3.2.2 Gender and life condition

All rays captured by the nets with control TEDs showed a slight sex-ratio bias towards females: smooth butterfly ray (57% female), longnose stingray (63%), smalleye roundray (51%) and sharpsnout stingray (64%) (Figure 11, Table 4). However, the only shark species, smalleye smooth-hound shark, showed a bias towards males with only 44% being female.

15 (a) (b)

(c) (d)

(e) (f)

Figure 10. Size frequency distributions of longnose stingray, smooth butterfly ray and smalleye smooth- hound shark taken as bycatch in nets fitted with control TEDs (white) and test TEDs (black).

16

Figure 11. Comparison of sex ratios (shown as percentage of females) for sharks and rays taken as bycatch in nets fitted with control and test TEDs. Red line signifies 1:1 female to male ratio.

For all three of the most commonly caught species: longnose stingray, smooth butterfly ray and smalleye smooth-hound shark, the size frequency distributions for those caught in the nets with control TEDs appeared similar between sexes (Figures 12-14), and there was no difference in mean size between sexes for the smalleye smooth-hound shark (Table 4). However, for the two ray species there were significant differences in mean disc width between the sexes for those caught in nets with control TEDs (Table 4).

Table 4. Summary of data and statistical results for a comparison between sexes of the mean sizes of shark and ray species taken as bycatch, shown separately for nets with control TEDs and test TEDs.

TED design Control Test Mean size Mean size Species Sex n t p n t p (cm) (cm) Longnose stingray M 21.09 115 18.35 20 2.461 0.542 0.590 0.014* F 19.40 198 18.00 33

Rays Smooth butterfly ray M 24.90 373 24.40 164 -5.116 <0.001* 0.742 0.459 F 27.82 487 23.99 163

M 22.03 60 20.29 7 Smalleye smooth- -0.715 0.476 -1.306 0.201 hound shark Shark F 22.63 48 22.28 25

* indicates statistical significance

17

Figure 12. Size frequency distribution of longnose stingray (shown separately by sex) taken as bycatch in nets fitted with control TEDs (white) and test TEDs (black).

18

Figure 13. Size frequency distribution of smooth butterfly rays by sex and 3 cm disc width size groups taken as bycatch in nets fitted with control TEDs (white) and test TEDs (black).

19

Figure 14. Size frequency distribution of smalleye smooth-hound shark by sex and 3 cm fork length size groups taken as bycatch in nets fitted with control TEDs (white) and test TEDs (black).

20 Almost all (91.4%) sharks and rays that were caught in the control nets were already dead upon being brought on board. However, of those that were caught alive, most were in good condition (7.6% of the total shark and ray bycatch), compared with 1.1% that were considered to be in poor condition and unlikely to survive when thrown back (Figure 15a).

Figure 15. Proportion of sharks and rays dead and alive in good or poor ‘life condition’ when landed from trawl nets fitted with (a) control TEDs and (b) test TEDs.

3.2.3 Catch rates

All sharks and rays

Over the duration of the study, an average of 2.00 (± 0.14) individual sharks and rays were caught as bycatch per standard unit of fishing effort (i.e. per hour of twin-net trawling) when using the control TEDs (Table 5). This standardised catch rate was relatively consistent across fishing trips; ranging between 1.77 and 2.50 sharks and rays per twin-net hr-1, and also showed low variation among trawls within a trip (Table 5).

Table 5. Summary of control TED shark and ray bycatch and mean standardised catch rates (no. individuals per twin-net hr-1) shown overall and separately for each trip (n = 182 twin-net hauls over 734.02 hr).

No. Duration Mean catch Trip No. caught SE twin-net hauls (twin-net hr) per twin-net hr-1 1 82 303.50 517 1.77 0.16 2 47 180.42 408 2.32 0.37 3 30 142.31 256 1.80 0.20 4 17 71.95 154 2.50 0.67 5 6 28.88 64 2.28 0.52 Total 182 734.02 1399 2.00 0.14

21 Commonly caught species

Smooth butterfly ray had the highest catch rate (1.22 ± 0.11 individuals per twin-net hr-1) (Table 6). The four remaining prominent shark and ray species were all caught at a rate of less than one individual hr-1: i.e. longnose stingray (0.46), smalleyed round stingray (0.06), sharpsnout stingray (0.06) and smalleye smooth-hound shark (0.15) (Table 6).

Table 6. Summary of control TED shark and ray bycatch rates (no. individuals per twin-net hr-1) shown separately for the commonly caught species (n = 182 twin-net hauls over 734.02 hr).

Species No. caught Mean catch per twin-net hr-1 SE Longnose stingray 313 0.46 0.04 Sharpsnout stingray1 44 0.06 0.02 Smalleyed round stingray 45 0.06 0.02 Smooth butterfly ray 860 1.22 0.11 Smalleye smooth-hound shark 108 0.15 0.02 All species 1399 2.00 0.14 1 indicates an IUCN Near Threatened species.

3.3 Shark and ray bycatch using modified test TEDs

3.3.1 Species composition and individual size

All sharks and rays

A total of 435 sharks and rays were caught across four orders, five families and six species in the trawl nets with the test TEDs (Table 2).

Again, most of the rays captured were small; the mean disc width for rays that passed through the test TED being 22.95 cm (± 0.28 cm) (Table 3), with the majority being between 19-21 and 22-24 cm wide (28% each) (Figure 8b). The smallest ray captured had a disc width of just 5 cm and the largest ray had a disc width of 55 cm. All specimens were wider than the test TED bar spacing (4.5 cm), and must therefore be passing through sideways or being folded and forced through presumably with significant injury. The only shark species is shown in the next section.

Commonly caught species

Again there were five shark and ray species which accounted for 99% of the elasmobranch bycatch in nets with test TEDs (Figure 9b). The most numerous species caught was the smooth butterfly ray, accounting for 75% of all individuals, whilst the others were: longnose stingray (12%), smalleye smooth-hound shark (7%), smalleyed round stingray (3%) and sharpsnout stingray (2%) (Figure 9b).

All four prominent ray bycatch species taken by the test nets had a mean width greater than the bar spacing of the test TED (4.45 cm): smooth butterfly ray (24.05 cm), longnose stingray (18.13 cm), smalleyed round stingray (11.45 cm) and sharpsnout stingray (28.43 cm) (Table 3). For the

22 smalleye smooth-hound shark, the mean length for individuals collected in the test nets was small at just 21.84 cm and the mean girth was 7.91cm, suggesting that the sharks are passing through the TED head on or at least parallel to the bars.

The size frequency distributions of the three most common species caught by the nets with test TEDs are shown in Figure 10. Longnose stingray were found between 14 and 26 cm, with 77% in the 16-20 cm size class (Figure 10b). No mature females, those ≥ 50 cm, were retained. For smooth butterfly ray, the size range was between 16 and 55 cm; half of this species was in the 25 cm size class (Figure 10d). Mature females, those ≥ 34 cm, comprised only 3.1% of the female smooth butterfly rays caught with the test TEDs. Smalleye smooth-hound shark were found to range between 17 and 37 cm long, with half in the 20 cm size class (Figure 10f).

3.3.2 Gender and life condition

Smooth butterfly rays captured by the test TEDs had a 1:1 sex ratio, whilst the other species were biased towards females; longnose stingray (62%), smalleyed round stingray (82%), sharpsnout stingray (71%) and smalleye smooth-hound shark (78%) (Figure 11). The size frequency distributions of males and females were very similar for each of the common elasmobranch bycatch species caught with the test TEDs (Figures 12-14) and there were no differences between the sexes in the mean sizes caught (t-tests, p < 0.001 in all cases, Table 4).

The bycatch of sharks and rays in the test TEDs consisted primarily of dead individuals (94.7%, Figure 15b). Sharks and rays found in good condition accounted for just 4.8% of the test TED elasmobranch bycatch, whilst those found alive but in poor condition accounted for 0.5% of the bycatch (Figure 15b).

3.3.3 Catch rates

All sharks and rays

A mean of 1.37 (± 0.22) individuals hr-1 of sharks and rays were caught while using the test TED (Table 7). Between 0.47 and 1.76 sharks and rays per twin-net haul were caught per hour of trawling, relatively consistent between fishing trips. There was little variation among trawls within a trip.

Table 7. Summary of test TED shark and ray bycatch and mean standardised catch rates (no. individuals per twin-net hr-1) shown overall and separately for each trip (n = 80 twin-net hauls over 329.6 hr).

No. twin-net Duration Mean catch per Trip No. caught SE hauls (twin-net hr) twin-net hr-1 2 47 180.42 315 1.76 0.35 3 10 48.35 42 0.87 0.20 4 17 71.95 65 0.92 0.32 5 6 28.88 13 0.47 0.11 Total 80 329.60 435 1.37 0.22

23 Commonly caught species

Smooth butterfly ray were caught at the highest mean rate (1.03 rays hr-1) (Table 8). The four remaining prominent shark and rays species were all caught at a rate less than one per hour: longnose stingray (0.17), smalleye smooth-hound shark (0.10), smalleyed round stingray (0.03) and sharpsnout stingray (0.02) (Table 8).

Table 8. Summary of test TED shark and ray bycatch rates (no. individuals per twin-net hr-1) shown separately for the commonly caught species (n = 80 twin-net hauls over 329.6 hr).

Species No. caught Mean no. caught per twin-net hr-1 SE Longnose stingray 53 0.17 0.03 Sharpsnout stingray1 7 0.02 0.02 Smalleyed round stingray 11 0.03 0.01 Smooth butterfly ray 327 1.03 0.20 Smalleye smooth-hound shark 32 0.10 0.02 All species 435 1.37 0.22

3.4 Comparison of shark and ray bycatch between control and test TEDs

3.4.1 Species composition and individual size

All sharks and rays

An overall total of 1,834 sharks and rays were caught as bycatch over the course of this study, comprising eight species of ray and one species of shark (Figure 7). All nine species were taken in the control TEDs and six species were taken in the test TEDs (Table 2). Three ray species that occurred in very small numbers in the control nets: cownose ray (0.6% of the shark and ray bycatch), chola guitarfish (0.4%) and southern stingray (0.4%) were not found at all in the test net bycatch (Table 9), even though these three species were caught during simultaneous trawls with control and test TEDs in operation.

The other six elasmobranch bycatch species were caught by both TED designs, and shared the same frequency ranking (Table 9). Smooth butterfly ray was by far the most commonly caught species constituting 61% of the control net shark and ray bycatch and 75% of the test net bycatch. Longnose stingray was the second-most abundant species caught, accounting for 22% of the control net bycatch and 12% of the test net bycatch (Figure 9). The third ranking species, the smalleye smooth-hound shark represented 8% and 7% of the bycatch of control and test TEDs respectively, whilst the other species accounted for less than 4% each in both TED designs (Table 9, Figure 9).

24 Table 9. Summary of shark and ray bycatch taken by control and test TEDs shown separately for each species as number, % frequency and frequency rank.

TED Control Test No. No. Species % Rank % Rank caught caught Bancroft's numbfish 10 0.7 6 5 1.1 6 Chola guitarfish 6 0.4 8 0 - - Cownose ray 8 0.6 7 0 - - Longnose stingray 313 22.4 2 53 12.2 2 Sharpsnout stingray 44 3.1 5 7 1.6 5 Smalleyed round stingray 45 3.2 4 11 2.5 4 Smooth butterfly ray 860 61.5 1 327 75.2 1 Southern stingray 5 0.4 9 0 - - Smalleye smooth-hound shark 108 7.7 3 32 7.4 3 Total 1399 100 435 100

The size frequency distributions of rays taken by the two TED designs appear similar for the main size groups, but the size range taken by the test TEDs is considerably reduced, with a virtual absence of the larger size classes (> 36 cm disc width) (Figure 8). Supporting this, the mean size of rays taken by test TEDs is significantly smaller (6.3% smaller) than for those taken by the control TEDs (t-test: t = 4.05, n = 1291, 404, p<0.001, Table 10). This pattern holds for both the commonly occurring ray species in the bycatch (longnose stingray and smooth butterfly ray), which show a statistically significant and 9.4% reduction in mean size when using the test TEDs (p <0.0001 in both cases; Figure 10, Table 10). For the shark (smalleye smooth-hound) there is a much smaller difference in the size range taken by the two different TEDs and although the mean size (length and girth) of shark is slightly smaller in the test TEDs (2.1% smaller length and 3.7% smaller girth), the differences are not significant (p >0.05 in both cases; Figure 10, Table 10).

Table 10. Comparison of mean size between shark and ray species taken as bycatch in nets with control TEDs versus test TEDs. Results shown are from two sample t-tests.

TED Control Test Comparison % smaller Species Width (cm) Width (cm) t-value p-value in test TED

Longnose stingray 20.02 18.13 9.4% 4.15 <0.001* Smooth butterfly ray 26.55 24.20 9.4% 5.97 <0.001*

Rays Total ray species 24.48 22.95 6.3% 4.05 <0.001* FL (cm) FL (cm)

22.30 21.84 2.1% 0.59 0.553 Smalleye smooth-hound Girth (cm) Girth (cm) Sharks 8.21 7.91 3.7% 0.92 0.180 * indicates statistical significance

25

The smaller bar spacing and horizontal grid of the test TEDs appears to have been successful in reducing the amount of larger rays which were able to pass through compared with the control TEDs.

3.4.2 Gender and life condition

Catches of all ray species showed a female bias in the nets with control TEDs, and an even greater bias towards females for some species taken by the nets with test TEDs (Figure 11, Table 11). However, the TED type only had a significant influence on the sex ratio of the catch for the most commonly caught ray species (smooth butterfly ray) and for the smalleye smooth-hound shark) (Table 11). Across all species, there was no significant difference in the sex ratio of catches between the two TED designs (p = 0.286).

Table 11. Summary data showing the proportion of male and female individuals caught by control TEDs versus test TEDs. Data are shown for all species together, and for the most common elasmobranch bycatch species. Significance levels are shown for 2x2 Pearson’s chi-square tests for independence.

No. caught Species Sex p-value Control TED Test TED Total Male 115 20 135 Longnose stingray 0.890 Female 198 33 231 Male 373 164 537 Smooth butterfly ray 0.036* Female 487 163 650 Smalleye smooth-hound Male 60 7 67 <0.001* shark Female 48 25 73 Male 593 197 790 All species 0.286 Female 806 238 1044 * indicates statistical significance

Most shark and rays that were caught in the industrial seabob fishery were found dead upon being brought on board, regardless of the TED type used, and of the few that were caught alive, most were found to be in good condition, and would likely survive if thrown back (Figure 15, Table 12). The TED type had no significant influence on this ratio (Table 12).

Table 12. Summary data showing the proportion of individuals in each ‘life condition’ caught by control TEDs versus test TEDs for all elasmobranch bycatch species. Significance levels are shown for 2x2 Pearson’s chi-square tests for independence.

No. caught Condition p-value Control TED Test TED Total Good 106 21 127 Poor 15 2 17 0.068 Dead 1278 412 1690

26 3.4.3 Catch rates

With simultaneous trawling using twin-nets fitted with control TEDs at the same time as twin- nets fitted with test TEDs, the overall catch rate of sharks and rays in the test TEDs was found to be significantly reduced (40.0%) from 2.289 to 1.374 individuals hr-1 (paired t-test: p<0.001, Table 13). Statistically significant declines in the catch rate were observed for three of the most commonly caught ray species: longnose stingray (66.1%, p<0.001), sharpsnout stingray (74.0%, p = 0.007) and smooth butterfly ray (28.4%, p<0.001, Table 13), and was also reduced for the only shark species, although not significant (27.4%, p = 0.276, Table 13). Of particular interest is the significant and very substantial reduction of bycatch for the IUCN listed near threatened species, the sharpsnout stingray.

Table 13. Comparison of mean standardised bycatch rates (no. individuals per twin-net per hour of trawling) for shark and ray species in simultaneous trawls (n = 80 trawls). Results are from paired t-tests.

Mean catch rate % reduction in Species TED (catch per twin-net SE t-value p-value catch rate hour) Control 0.493 0.071 Longnose stingray 66.1% 5.01 <0.001* Test 0.167 0.034 Control 0.088 0.029 Sharpsnout stingray1 74.0% 2.77 0.007* Test 0.023 0.015 Smalleyed round Control 0.061 0.019 47.7% 1.87 0.065 stingray Test 0.032 0.012 Control 1.448 0.214 Smooth butterfly ray 28.4% 3.79 <0.001* Test 1.037 0.202 Smalleye smooth-hound Control 0.139 0.033 27.4% 1.10 0.276 shark Test 0.101 0.021 Control 2.289 0.264 All species 40.0% 5.55 <0.001* Test 1.374 0.220 * indicates statistical significance. 1 indicates IUCN Near Threatened species.

27 4 DISCUSSION

This study represents the first documentation of the elasmobranch bycatch in the Guyanese industrial seabob fishery, and also examines the efficacy of a modified TED design in further reducing the bycatch of sharks and rays as part of a fisheries improvement programme towards attaining MSC certification of the seabob fishery in Guyana.

The elasmobranch bycatch in the industrial seabob trawl fishery, as it currently operates with all nets fitted with a BRD (10 x 10 mesh square of 10 cm mesh size) and a TED (with 8.9 – 10.2 cm grill spacing), comprised one shark and eight ray species. The catch was dominated by two mid- sized rays: smooth butterfly ray and longnose stingray, both listed as ‘Data Deficient’ by the IUCN Red List (http://www.iucnredlist.org/). These two species alone accounted for 83% by number of the elasmobranch bycatch. A further 8% of the catch comprised the small, relatively abundant bottom-dwelling shark, the smalleye smooth-hound, considered by the IUCN Red List as being of ‘Least Concern’. All other ray species represented 3% or less of the catch. However, of the two small-sized rays (smalleyed round stingray and Bancroft’s numbfish), the latter, representing 1% of the catch, is considered ‘Critically Endangered’ by the IUCN Red List (de Carvalho, McCord and Myers 2007). Furthermore, three of the mid- to large-sized rays taken as a small component of the bycatch (cownose ray [0.6%], sharpsnout stingray [3%], chola guitarfish [< 0.5%]) are listed as ‘Near Threatened’ (see Barker 2006, Charvet-Almeida and Almeida 2006, and Casper and Burgess 2009 respectively). This highlights the importance of further reducing the elasmobranch bycatch if the fishery is to be considered sustainable over the long-term, based on the MSC standards.

Interestingly, the industrial seabob fishery in neighbouring Suriname, which has also incorporated very similar BRDs (11 x 11 mesh square of 15 cm-stretched mesh) and TEDs (with 10 cm grill spacing) in all of their trawl nets, shares the same key ray bycatch species composition according to a year-round study with a similar sample size (n = 1,229 ray specimens) to our own study (n = 1,291) (Willems et al. 2016). For example, the four most abundant ray species in the bycatch of both fisheries share the same importance rank (assigned by relative number caught) and account for 97.8% of the Guyanese ray bycatch and 99.3% of the Suriname ray bycatch (Willems et al. 2016). The only difference lies with the rarely caught species, of which there is just one in the Suriname fishery (cownose ray) and an additional three recorded in our own study for the Guyana fishery (Bancroft’s numbfish, chola guitarfish and southern stingray).

The mean size of rays caught in our study with the control TEDs (24.5 cm disc width) is slightly smaller than that reported in the Surinamese study (25.5 cm disc width) (Willems et al. 2016). The most obvious explanation for this slightly smaller size in our Guyana study is the use of some TEDs with a slightly smaller grid spacing than in the Surinamese study (i.e. 8.9 vs 10 cm). This corroborates the information that these two fisheries operating in adjacent areas along the coast of Guyana and Suriname are very similar in terms of their gear, fishing operations and fishing ground habitats (Maison 2016). It also suggests that modifications made to the gear in Guyana is likely to have a similar impact on the elasmobranch bycatch if also implemented in the Suriname seabob fishery.

28 In the current study, there was no evidence that the modified TED design had any real effect on the elasmobranch bycatch species composition compared with the control TED currently used throughout the fishery. For example, the six elasmobranch species caught by the test TED design (1 shark and 5 rays) accounted for 98.6% of the catch of the control TED design and each species shared the exact same abundance rank across both TEDs. Although three species were not caught at all in the nets with test TEDs over the course of the study, they were caught in such low abundance in the control nets (together representing < 1.4% of total elasmobranch bycatch) that their absence in the test nets does not provide convincing evidence that these TEDs consistently released them. Furthermore, there was no evidence that the use of the test TED altered the ‘life condition’ of the bycatch.

Of great significance is the fact that the use of the test TED did affect the mean size of the elasmobranchs taken as bycatch, by allowing the larger specimens to pass out of the nets, and most importantly, the use of the test TED resulted in significant reductions in the overall catch rates of elasmobranchs. These results can indeed be attributed to the differing TED design, since the BRD dimensions remained the same across both net types.

Overall, the mean size (disc width) for the ray bycatch was reduced by 6.3% in the nets fitted with the test TED compared with the control TED nets, although reductions in size were not the same across all species. The test TEDs effectively eliminated the capture of any rays with a disc width > 36 cm. As such, highly significant reductions in mean size were observed for both the dominant mid-sized ray bycatch species: smooth butterfly ray and longnose stingray (both reduced by 9.4%), and a 12.5% reduction in size was noted for the mid-sized sharpsnout stingray, although not tested statistically because of a relatively small sample size. By contrast, the mean size of the small-sized ray species caught by both nets, the smalleyed round stingray, was actually larger by 22.3% in the test TEDs, although again was not tested for significance due to a very small sample size. It is not surprising that a reduction in the size of the TED grill spacing had the greatest success in preferential exclusion of larger individuals of the larger-sized ray species, whilst most of the larger individuals of the small-sized species that were retained in the control TED nets were still able to pass through the grill into the retained catch when using the test TEDs. The smaller the individual, the more likely it is to pass through sideways, or fold and get forced through the TED grill. Since the maximum size of the smalleyed round stingray is just 12 cm disc width (Uyeno, Matsuura and Fujii 1983, cited in Fishbase), it can be expected that most individuals would be small enough to pass or be forced through the modified grill (4.5 cm) and be captured. With regard to the single species shark bycatch, the test TED did not significantly reduce the mean size caught, although they were generally smaller (by 2.1%). Given the fusiform shape and small size of the smalleye smooth-hound shark species, it is again not surprising that even a substantial reduction in grill spacing width (from 10.2 to 4.5 cm) was insufficient to prevent their capture. For example, the mean girth of the sharks retained by the control TED (8.21 cm) would suggest a mean diameter < 2.6 cm, i.e. smaller than even the test TED grill spacing, thus making little difference to the capture size.

More surprising was the female biased sex ratio of the mid-sized ray bycatch, and the apparent effect of the test TEDs in decreasing that bias for the most commonly caught species, the smooth butterfly ray. Furthermore the shark retained by the control TED had a slight male bias, but those

29 retained by the test TED were almost all female. These significant gender biases could suggest growth rate/size differences between the sexes or behavioural differences and/or spatial segregation of the sexes, resulting in greater vulnerability of one or the other sex to the fishing gear. However, it does not explain how the narrower grid spacing altered the sex ratio bias.

Another important result for the fishery is the near elimination of mature females in the ray bycatch of the test TEDs. Using published size-at-first maturity data for females of the two most abundant ray species (smooth butterfly ray: 34-36 cm; longnose stingray: 50-55 cm, Yokota and Lessa 2007) we find a substantial reduction from 10.9% to just 3.1% of mature females for the former species and a total elimination (from 0.5% to 0%) for the later species. Given the concomitant reduction in the catch rate also, this represents an even more significant reduction in the capture of mature females. Like other fish, elasmobranchii fecundity increases exponentially with size (Stevens et al. 2000), but compared with teleost fishes their fecundity is extremely low, such that it is especially important for larger females to remain in the breeding stock.

The overall capture rate for all elasmobranchs combined fell by 40% when using the test TEDs compared with the control TEDs, which is an enormously important result for the Guyanese seabob fishery. As with the effect on mean size caught, the change in capture rate varied among species. Most importantly the capture rate was significantly reduced for the two dominant species in the bycatch, longnose stingray (by 66.1%) and smooth butterfly ray (by 28.4%). Also of importance was the substantial reduction in catch rate of the sharpsnout stingray (by 74.0%) since it is listed by the IUCN as ‘Near Threatened’ due to its limited habitat range and its frequent capture by industrial fisheries (Charvet-Almeida and Almeida 2006). Not surprising, was the lack of significance in the observed reductions in catch rate for the very small-sized species: the smalleyed round stingray, and the smalleye smooth-hound shark. The Suriname study even reported no significant effect of using a TED versus no TED on the mean size of the smalleyed round stingray captured, citing their small size as the reason (Willems et al. 2016).

Of interest is the marked difference in the magnitude of catch reductions for the two similar- sized dominant ray species, with the longnose stingray showing a much greater reduction in catch rate than for the smooth butterfly ray. This difference was also reported in the Suriname study by Willems et al. (2016) when comparing the catch rates of these two species in nets with and without TEDs. They suggested that this finding resulted from a difference in the morphology of the two species, with longnose stingrays having a thicker and more rigid disc, whilst the smooth butterfly ray was thinner and more flexible, and thus more easily distorted and forced through the TED grill.

Also of interest is a comparison of the bycatch rates between our study in Guyana and those reported in Suriname when using virtually the same BRD and TEDs and standardising catch rates to number of individuals caught per twin-net hour. Our own ray bycatch rate with control TEDs is an order of magnitude less than that reported for the Suriname fishery when using TEDs (Guyana: 1.9 rays hr-1 ± 0.17 SE; Suriname: 15.3 rays hr-1 ± 1.60 SE, Willems et al. 2016). There are several plausible explanations including: 1) the density of rays is much greater on the Suriname fishing grounds, 2) the density of rays is much lower during the summer, and/or 3) rays are less susceptible to capture during night tows. The most likely is that the density of rays

30 are higher on the Suriname fishing grounds compared with Guyana possibly as a result of lower fishing pressure, given the smaller seabob trawl fleet size (20 vessels cf. 88 Guyanese vessels) but similar delimited fishing area, and the fact that the Suriname fishery is also a much younger fishery (started around 1996 cf. mid-1980s for the Guyanese seabob fishery; see Willems et al. 2013). The catch rate of rays did show variation between months in the Suriname study that took place over the period of 14 months, but there was no obvious seasonal pattern and the lowest catch rates did not occur during the summer months (Willems et al. 2013). Although the Suriname study only fished during daytime, we found no difference in our catch rates between day and night hauls.

Based on these results, including our shipboard observations and a review of the technical literature on bycatch reduction, we have considered ways in which the efficacy of the test TED could perhaps be further improved to release the small-sized elasmobranchs. We suspect that a further reduction in horizontal bar spacing may lead to a decline in the catch rate of seabob that would be unsustainable economically. However, there are several alternatives that may be worth trying. Firstly, the addition of one or two more horizontal brace or cross bars to the test TED may prevent the small-sized and flexible individuals, especially the rays from being forced through the grill. Secondly, the industry currently uses oval-shaped TEDs, which are effective at retaining net shape while minimising stress and abrasion on the net itself. Rectangular-shaped TEDs have been reported to be more effective at reducing bycatch of wider animals, as it allows more room for them to manoeuvre through the escape flap (Eayrs and Day 2004, Eayrs 2007). However, it is acknowledged that rectangular-shaped TEDs, cause more abrasion of the net, leading to a decrease in net condition and TED efficiency (Eayrs and Day 2004). A compromise could be to use a hybrid of the two designs; i.e. a tombstone-shaped TED which would provide greater net width to accommodate shark and ray species attempting to escape, while also reducing abrasion on the net. Another approach could be to focus on the TED grid orientation. The TEDs currently being used in the Guyanese industrial trawl fishery are ‘bottom exclusion’, in that large bycatch species escape through the bottom of the net. Eayrs and Day (2004) have suggested that top exclusion may be more effective in reducing mobile bycatch, whilst bottom exclusion TEDs are more effective at removing rubble and sponges. As no rubble or sponges were spotted in the cod-ends during this study, converting to a top exclusion TED may further limit the bycatch of sharks and rays and indeed of finfish. We recognise however, that none of these TED ‘improvements’, including the GAPTO&SP modified test TED that we examined should be slated for mandatory adoption without an examination of the impact on the target seabob species, which was beyond the scope and feasibility of the current study.

31 5 CONCLUSION

This represents the first study to examine elasmobranch bycatch in the improved industrial seabob trawl fishery of Guyana with the compulsory use of standard BRDs and TEDs, and to examine the efficacy of a modified TED, with substantially reduced grill spacing, in further reducing elasmobranch bycatch. The most important results of this study for the fishery, given the interest in becoming MSC certified are:

 A total of eight ray species and one shark species are currently taken as bycatch in the industrial seabob fishery in Guyana.

 The most commonly caught species comprise two rays, both considered to be ‘Data Deficient’ by the IUCN Red List, and a small shark considered to be of ‘Least Concern’.

 Among the less commonly caught species are three rays listed as ‘Near Threatened’ by the IUCN Red List, and one listed as ‘Critically Endangered’.

 The use of the test TED did not affect the key elasmobranch species composition in the bycatch.

 The use of the test TED resulted in a significant reduction (by 40%) in bycatch rate of elasmobranchs.

 The use of the test TED was highly effective in reducing the capture of the three ‘Near Threatened’ ray species, with a very significant reduction (by 74%) in the capture for one of them and a complete absence of capture for the other two species.

 The preferential release of larger-sized individuals by the test TED resulted in a significant reduction (by 6.3%) in the mean individual size of elasmobranchs taken as bycatch, essentially eliminating the capture of rays larger than 36 cm disc width.

 The test TED resulted in the near elimination of mature females of the two most commonly caught ray species in the bycatch.

 The test TED was most effective at reducing the capture rate and mean size of retained individuals for the larger-sized species, and made little difference for the infrequently caught smallest species, including the ‘Critically Endangered’ ray.

 The Guyanese industrial seabob fishery has many similarities with the MSC certified Suriname industrial seabob fishery, including the same fishing practices and key species composition of the ray bycatch. Furthermore, the Guyanese fishery appears to have a substantially lower ray bycatch rate compared with the Suriname fishery, and this would be further reduced by the mandatory adoption of the modified test TED.

32  Given the similarities between the two national seabob fisheries, the impacts of the modified TEDs observed in this study in Guyana are likely to be entirely applicable to the Suriname fishery also.

Coming out of the results of the present study, we can also make some suggestions for further informative research to assist in the continued development of an MSC certifiable seabob fishery in Guyana. These include:

 An extension of the current study to span a full year, to give more confidence in the estimates of impacts across the entire fishery.

 An investigation of the impacts of the modified test TED on the target seabob catch rate.

 Expansion of trials with other slight modifications to the test TED including for example the overall TED shape, setting angle within the net, addition of several more horizontal bars, to determine whether this increases the efficacy, especially with regard to the small- sized elasmobranch species.

33 6 REFERENCES

Arjune, Z., S. Richardson, D. Maison and P. Medley. 2013. Report of the Shrimp and Groundfish Resource Working Group (SGWG). In: Report of the Ninth Annual CRFM Scientific Meeting, 43-44. Belize: CRFM. Barker, A. 2006. Rhinoptera bonasus. Accessed September 19, 2014 www.iucnredlist.org/details/60128/0. Bonfil, R. 1994. Overview of world elasmobranch fisheries. FAO Fisheries Technical Paper 341. Rome: FAO. Brewer, D., D. Heales, D. Milton, Q. Dell, G. Fry, B. Venables and P. Jones. 2006. The impact of turtle excluder devices and bycatch reduction devices on diverse tropical marine communities in Australia's northern prawn trawl fishery. Fish Res. 81: 176-188. Camhi, M., S. Fowler, J. Musick, A. Bräutigam and S. Fordham. 1998. Sharks and their Relatives – Ecology and Conservation. Gland, Switzerland and Cambridge, UK: IUCN. Casper, B. and G. Burgess. 2009. Rhinobatos percellens. Accessed September 19, 2014 www.iucnredlist.org/details/161373/0. Charvet-Almeida, P. and M. de Almeida. 2006. Dasyatis geijskesi. Accessed September 19, 2014 www.iucnredlist.org/details/60153/0. Compagno, L.V.J. 2002. Sharks. In: The Living Marine Resources of the Western Central Atlantic, ed. K.E. Carpenter, 358-505. Rome: FAO. Cox, T., R. Lewison, R. Zydelis, L. Crowder, C. Safina and A. Read. 2007. Comparing effectiveness of experimental and implemented bycatch reduction measures: The ideal and the real. Conservation Biology 21: 1155-1164. Crespi, V. and J. Prado. 2002. Fishing technology equipments. Bycatch Reduction Device (BRD). Accessed June 21 http://www.fao.org/fishery/equipment/brd/en. Crespi, V. and J. Prado. 2002. Fishing Technology Equipments. Turtle Excluder Device (TED). Accessed June 21 http://www.fao.org/fishery/equipment/ted/en. CRFM. 2007. Report of the Third Annual Scientific Meeting. CRFM Fishery Report 2007, Vol. 1. Belize: CRFM. CRFM. 2013. Report of Ninth Annual CRFM Scientific Meeting. CRFM Fishery Report 2013, Vol. 1 Belize: CRFM. de Carvalho, M., M. McCord and R. Myers. 2007. Narcine bancroftii. Accessed September 19, 2014 www.iucnredlist.org/details/63142/0. Earl, S. 2009. Venezuela Bans Controversial “Trawl” Fishing. Accessed April 11, 2016 http://venezuelanalysis.com/news/4302 Eayrs, S. 2007. A guide to bycatch reduction in tropical shrimp-trawl fisheries. Revised edition. Rome: FAO. Eayrs, S. and G. Day. 2004 Optimising TED performance in tropical prawn-trawl fisheries: a review. Prof. Fisherman 26(8), 16-18. European Commission. 2014. Illegal Fishing (IUU). Accessed December 11, 2014 http://ec.europa.eu/fisheries/cfp/illegal_fishing/index_en.htm. FAO. 2003. The ecosystem approach to fisheries. FAO Technical Guidelines for Responsible Fisheries No. 4, Suppl. 2. Rome: FAO. FAO. 2005. Fishery country profile – the Repulic of Guyana. http://www.fao.org/fi/oldsite/FCP/en/GUY/profile.htm

34 FAO. 2011. International guidelines on bycatch management and reduction of discards. Rome: FAO. Food Certification International Ltd (FCI). 2011. MSC Sustainable Fisheries Certification: Off- Site Surveillance Visit – Report for the Suriname Atlantic sea-bob shrimp Fishery. Accessed June 21 http://www.msc.org/track-a-fishery/fisheries-in-the- program/certified/western-central-atlantic/suriname_atlantic_seabob_shrimp/assessment- downloads-1/Suriname_Seabob_PCR_Nov2011.pdf. Hall, M. 1996. On bycatches. Reviews in Fish Biology and Fisheries 6: 319-352. IUCN. 2014. A quarter of sharks and rays threatened with extinction. Accessed December 11 www.iucn.org/?14311/A-quarter-sharks-and-rays-threatened-with-extinction. Holthuis, L.B. 1980. FAO Catalogue Vol. 1 – Shrimps and Prawns of the World. An Annotated Catalogue of Species of Interest to Fisheries. FAO Fisheries Synopsis 125. Rome: FAO. Kalicharan, L. In prep. Examining finfish bycatch discards in the artisanal Chinese seine fishery in the Demerara estuary, Guyana. Research Paper, Centre for Resource Management and Environmental Studies, University of the West Indies, Cave Hill, Barbados. MacDonald, J., S. Harper, S. Booth and D. Zeller. 2015. Guyana fisheries catches: 1950-2010. Working Paper # 2015 – 21, Sea Around Us, Fisheries Centre, University of British Columbia, 18pp. Maison, D. 2007. Management of inshore artisanal fisheries in Guyana: A co-management approach. Reykjavik, Iceland: The United Nations University Fisheries Training Project research paper. Maison, D. 2016. Fishery improvement project: Guyana Atlantic seabob. FIP Directory, Accessed December 20, 2016, https://fisheryimprovementprojects.org/fip/guyana- atlantic-seabob/ McEachran, J.D., and M.R. de Carvalho. 2002. Batoid Fishes. In: The Living Marine Resources of the Western Central Atlantic, ed. by K.E. Carpenter, 507-589. Rome: FAO. Medley, P. 2014. Guyana HCR 2013/14. Peebles, UK: Nautilus Consultants. MSC. 2014. Updated fishery certification requirements v2.0. Accessed October 19, 2014 https://www.msc.org/about-us/standards/fisheries-standard/certification- requirements/fishery-certification-requirements-v2.0 NMFS. 2014. Shrimp import legislation for sea turtle conservation. Accessed December 11, 2014, www.nmfs.noaa.gov/pr/species/turtles/shrimp.htm. Richardson, S. 2013. “National Report of Guyana.” In: Report of the Ninth Annual CRFM Scientific Meeting: Supplement 1, 23-28. Belize: CRFM. Sala, A., A. Lucchetti, A. and M. Affronte, 2011. Effects of Turtle Excluder Devices on bycatch and discard reduction in the demersal fisheries of Mediterranean Sea. Aquatic Living Resources 24: 183-192. SOFRECO. 2013. Support to formulate fisheries management plans for Guyana, Suriname and Trinidad and Tobago. Accessed June 21, 2014 http://acpfish2-eu.org/uploads/projects/id146/FTR_new.pdf Stevens, J., R. Bonfil, N. Dulvy and P. Walker. 2000. The effects of fishing on sharks, rays, and chimaeras (chondrichthyans), and the implications for marine ecosystems. Ices Journal of Marine Science 57: 476-494. Stobutzki, I., M. Miller, D. Heales and D. Brewer. 2002. Sustainability of elasmobranchs

35 caught as bycatch in a tropical prawn (shrimp) trawl fishery. Fishery Bulletin 100: 800- 821. Taveres, M. 2002. Shrimps. In: The Living Marine Resources of the Western Central Atlantic, ed. K.E. Carpenter, 278. Rome: FAO. WECAFC. 2013. Investing in ecosystem-based shrimp and groundfish fisheries management of the Guianas-Brazil shelf. Accessed June 21, 2014 ftp://ftp.fao.org/FI/DOCUMENT/wecafc/15thsess/ref14e.pdf . Willems, T., J. Depestele, A. De Backer and K. Hostens. 2013. Bycatch of rays in the trawl fishery for Atlantic seabob shrimp Xiphpenaeus kroyeri in Suriname: how effective are TEDs and BRDs? Instituut voor Landbouw- en Visserijonderzoek 139. WWF. 2011. Making fishing sustainable. Accessed December 11, 2014 http://wwf.panda.org/?199902/Making-fishing-sustainable. Yokota, L., and R.P. Lessa. 2007. Reproductive biology of three ray species: Gymnura micrura (Bloch & Schneider, 1801), Dasyatis guttata (Bloch & Schneider, 1801) and Dasyatis marianae (Gomes, Rosa & Gadig, 2000) caught by artisanal fisheries in Northeastern Brazil. Cahiers de Biologie Marine 48: 249-257.

36