Efficiency of the bottom trawl used for the Mediterranean international trawl survey (MEDITS)

The aim of the work presented in this paper was to assess the relative efficiency of the GOC 73 sampling trawl used for the Mediterranean international trawl survey programme (MEDITS) compared with that of a typical Italian commercial trawl. The latter was chosen because it is commonly used by fishermen in the central Adriatic, where the experiment was conducted, and therefore appeared as the best possible sampler of the commercial species present in those areas. Moreover, this trawl is similar to the various trawls used for the Italian national survey programme (GRUND). Four fishing trips were conducted between 1996 and 1997 in different areas and seasons to sample different species. Each trip was conducted in the same fishing area. A codend cover mounted on the commercial trawl ensured that the codends of the two nets had the same mesh opening. The two trawls were alternated daily. Net geometry (horizontal and vertical net opening) was measured during all tows. Catch data were converted into abundance per swept area before comparing the trawls. Efficiency comparisons were performed on ten fish species, one crustacean and four molluscs belonging to the MEDITS list of main reference species. The MEDITS trawl was significantly less efficient in terms of both weight and numbers of individuals fished for hake (Merluccius merluccius), common sole (Solea vulgaris) and Norway lobster (Nephrops norvegicus). A highly significant difference in favour of the commercial trawl was found in the weight, but not the numbers, of common pandora (Pagellus erythrinus). Efficiency differences were negligible for red mullet (Mullus barbatus), while the MEDITS trawl was significantly more efficient for the numbers of Atlantic horse mackerel (Trachurus trachurus). For all the other main reference species differences were not significant. The commercial trawl was more efficient for large individuals of some species. The relative efficiency of the MEDITS trawl was especially low for small-size classes of N. norvegicus. The results confirm the lower efficiency of the MEDITS compared with the commercial Italian trawl for benthic species, and its greater efficiency for some others released from the bottom as well as for the pelagic ones. These data allow the results of the MEDITS surveys to be compared with those of the GRUND programme and with commercial-fleet catches in areas where the commercial trawls used are similar to the one studied here.     

Size selection by diamond- and square-mesh codends in multi-species Mediterranean demersal trawl fisheries

Mediterranean demersal trawl fisheries traditionally operate using small diamond-shape meshes in the codend, which tend to retain almost all animals. We investigated the effect of mesh configuration on the size selectivity of nine species (the scaldfish Arnoglossus laterna, the broad-tail shortfin squid Illex coindettii, the European hake Merluccius merluccius, the red mullet Mullus barbatus, the Norway lobster Nephrops norvegicus, the common pandora Pagellus erythrinus, the deepwater rose shrimp Parapenaeus longirostris, the Mediterranean horse mackerel Trachurus mediterraneus and the poor-cod Trisopterus minutus capelanus) commonly captured in the Mediterranean demersal trawls. Two codends having meshes with the same opening (ca. 38 mm) but different mesh configuration (DM38: diamond-mesh and SM38: square-mesh configuration) were fished daily and alternately on the same trawl. They demonstrated a substantial improvement in selectivity with square-mesh. Selectivity was measured using the covered codend technique, the cover being supported by circular hoops.The results were analysed taking into account the between-haul variation in selectivity. Two other important external variables were identified: the trawling depth and the codend catch, which for some species influenced between-haul variation, but there was no evidence of a coherent effect on selectivity parameters.The selectivity of 40 mm diamond-mesh codend has been reported to be rather poor because a large proportion of the codend catch is immature and smaller than the minimum landing size (MLS) or first maturity size. With the exception of the flatfish (A. laterna), the effect of a change of mesh configuration from diamond- to square-mesh on size selectivity positively affected the retention length at 50% (L₅₀). However, square-meshes were found to be unsuitable for flat and/or deep-bodied fish as these escape more readily from diamond-meshes.The Council Regulation (EC) No. 1967/2006, concerning management measures for the sustainable exploitation of fishery resources in the Mediterranean, establishes the MLS of marine organisms. However, the increase in L₅₀ with square-mesh would not avoid some of the existing contradictions in allowing the use of codend mesh which leads to lower L₅₀ than the MLS.     

Improving the effectiveness of escape windows in directed Norway lobster Nephrops norvegicus trawl fisheries

A substantial improvement in the bycatch selectivity of Norway lobster Nephrops norvegicus trawls is required, particularly with respect to cod Gadus morhua, whose stocks are at low levels in several areas. Conventional escape windows are not adequate to properly release cod and other bycatch species caught in the trawls. To address this issue, we developed a novel sorting box concept consisting of a four-panel section with a window on the top in order to improve the escape of cod and other bycatch species through an escape window while retaining the target catch of Norway lobster. The concept was tested on a commercial trawler in Kattegat and Skagerrak. Two different window mesh sizes and two different sorting box heights were tested using a traditional codend cover and a dual codend cover. We observed greatly reduced bycatches of both cod and other fish species compared to a standard codend. The reduction in bycatch decreased with decreasing mesh size and increasing height of the sorting box. Escape of Norway lobster through the escape window was limited. A modified version of the sorting box concept was implemented in the Kattegat fishery from 2009 onwards.     

The potential for new Nephrops trawl designs to positively effect North Sea stocks of cod,haddock and whiting

Within commercial fisheries, particularly mixed fisheries, both target and non-target species are often discarded. Discarding represents a potentially significant loss to the productivity of fish stocks; it can have damaging ecological consequences, and is a potential cause of the failure of recovery plans. The Nephrops fishery in the North Sea is classified as a mixed fishery. Nephrops trawls are constructed with smaller meshes than trawls used to target whitefish; consequently, the bycatch of juvenile fish can be substantial. Several new Nephrops trawl designs have been tested in the North Sea. The data from these trials are used to investigate the potential impact of their implementation on cod, haddock and whiting stocks in the North Sea (including the Kattegat and Skagerrak).The model examines five trawl designs, and also the scenarios of a cessation of discarding in all North Sea fisheries and in just the Nephrops fishery. The model is deterministic, and evaluates the relative differences between scenarios assuming all other variables remain constant. If discarding of cod, haddock and whiting in the North Sea fisheries were eliminated, stocks would increase by 41%, 14% and 29%, respectively, within 10 years. Eliminating discarding in the Nephrops fishery alone would increase stocks by 2%, 1% and 13%, respectively, reflecting the relative proportion of catches of these species in the Nephrops fishery. For cod and haddock, the introduction of the Nephrops trawl with a grid with a square-mesh codend was the only scenario in which a notable increase in stock number was observed. This trawl design facilitates the escape of fish of all ages/sizes from the trawl, effectively making the Nephrops fishery a single-species fishery. For whiting, stock numbers and landings increased under all scenarios, but forecasted landings were lower than if current discard patterns continued in all except the no-discards scenario. The dependency of the results on the validity of the assumptions and on the accuracy of the input data is discussed.     

Optimising codend configuration in a multispecies demersal trawl fishery

Abstract The relative efficiencies and selectivities of conventional and modified codends were examined in a demersal trawl fishery targeting several species, including eastern king prawns, Penaeus plebejus (Hess), whiting, Sillago spp. and cephalopods. The modifications to codends included: (i) reducing the circumference; (ii) increasing the mesh size in conventional diamond‐mesh designs; and (iii) orientating meshes on the bar so that they were square shaped. The codends were tested against a fine‐meshed control in paired comparisons onboard three commercial trawlers. The conventional codend comprised 41‐mm diamond‐shaped mesh attached to an anterior extension section at a ratio of 150 to 100 meshes and was demonstrated to be non‐selective for the targeted species. Reducing codend circumference to 100 meshes and increasing the size of mesh to 45 mm both improved selection for eastern king prawns, but the lateral mesh openings were estimated to be insufficient to allow juveniles of the other key species to escape. By contrast, codends made from 35‐ and 41‐mm mesh hung on the bar improved the size selection for eastern king prawns and selected stout whiting, Sillago robusta (Stead) (the smallest commercial‐sized fish) across narrow selection ranges and at 50% sizes of retention (L₅₀s) that were closely correlated to the transverse morphology of fish and the maximum mesh opening. With the exception of a reduction in catches of octopus, Octopus spp., by the 41‐square codend, there were no other impacts on commercial catches by the square‐mesh designs. It was concluded that diamond‐mesh codends are inappropriate for use throughout this multispecies fishery and that a modified design comprising at least 35‐mm mesh hung on the bar is required to minimise the fishing mortality of unwanted sizes of the key target species. The results are discussed in terms of the utility of these types of modifications for closely regulating selection in penaeid‐shrimp trawls.     

Selectivity in a trawl codend during haul-back operation—An overlooked phenomenon

The selectivity of a 99 mm trawl codend was assessed using a codend cover fitted with a MultiSampler, which was acoustically triggered to take separate samples at three different phases of the haul. The first sample was collected during towing, the second during haul-up and the third at the surface. A total of 18 hauls were conducted with a commercial fishing vessel west of Scotland. It was demonstrated that escapes take place not only during the tow but also in the short period when the trawl is hauled back from the seabed and when the codend is at the surface. For haddock (Melanogrammus aeglefinus), whiting (Merlangius merlangus) and Norway lobster (Nephrops norvegicus), respectively, the mean percentages escaping at the surface were 16, 12 and 38% of the total escape while 17, 8 and 28% escaped during the haul-up phase. Compared to towing, the escape rate (no./min) increased for haddock by a factor 2.7 during haul-up and by a factor 1.7 at the surface, whereas the escape rates of whiting were similar for the three phases. The escape rate of Norway lobster increased by a factor of approximately 7 for both the haul-up and surface phases, compared to the towing phase. The selectivity parameters L50 (50% retention length) and SR (selection range = L75–L25) were estimated and compared for the three different phases and for the whole haul for haddock, whiting and Norway lobster. For all three species there was no significant (P > 0.05) difference in L50 between the three phases of the haul. There was also no significant difference for whiting and Norway lobster when comparing the SR of the three phases, whereas the SR was significantly lower for haddock when comparing the surface phase with towing and haul-up. The estimate of L50 when towing was about 6 cm lower for haddock and whiting and 9 mm for Norway lobster compared to the selection curve estimated conventionally for the whole haul. Finally, the effect of sea state, duration and codend catch on the selectivity parameters were estimated for the individual phases and for the whole haul. A significant effect of at least one variable was found in all phases.     

Analysis of demersal species assemblages from trawl surveys in the South Adriatic sea

Two trawl surveys were carried out in the summer months (June–July) of 1996 and 1997 along the whole South Adriaticarea (Mediterranean sea) for the first time, using the same vessel and the same sampling gear (European Community Research Project `MEDITS'); previous research data referred only to the south-western Adriatic side. A total list of 168 demersal species (fishes, cephalopods and crustaceans) was obtained during the surveys; species abundance data (individuals/trawling hour) were processed accordingto multivariate techniques in order to describe the composition and the distribution of the main species assemblages within the investigated area (10–800-m bathymetric range). Multivariate analysis of MEDITS survey catch data showed a strong relationship betweenfish assemblages and depth, while the influence of depth was lower with respect to cephalopod and especially crustacean assemblage distribution.     

Codend selectivity for jack mackerel and whitefin jack and unequal split parameter estimates observed in trouser trawl experiments

Codend selectivity for the jack mackerel Trachurus japonicus and the whitefin jack Kaiwarinus equula were evaluated based on data from trouser trawl experiments carried out in the East China Sea, using a test codend of 60 mm diamond mesh and a control codend made of minnow net with a square mesh of 9 mm bar length. Between-haul variations in parameters and the mean selection curves were tested with the catch data in the SELECT approach, and then the model of between-haul variation in the split parameter with the mean selection curve was chosen as the best fit using Akaike’s information criterion model selection. The 50% retention lengths and the selection ranges were 11.4 and 3.36 cm for jack mackerel and 8.83 and 0.93 cm for whitefin jack, respectively. The selection curve for whitefin jack was sharp, whereas that for jack mackerel was relatively wide. As the estimated split parameters indicated, about 80% of the whitefin jack entered the control codend, but 85 and 90% of the jack mackerel entered the control codend in the second and third hauls, respectively. The inequality in the split parameter is discussed from the viewpoint of the animal’s swimming behavior and water movement based on underwater video observations.     

Selectivity and escape percentages during three phases of the towing process for codends fitted with different selection systems

We investigated the selectivity of cod (Gadus morhua) and haddock (Melanogrammus aeglefinus) escaping from three different selection systems (a conventional diamond-mesh codend, a codend fitted with escape windows, and a codend fitted with a sorting grid) using the covered codend technique. By using a MultiSampler attached to the codend cover to partition the collection of escaping fish, we were able to determine the partitioned selectivity of each of these devices at different phases during the towing process: at depth, during the haul back, and at the surface.The percentages of escaping fish at depth from the diamond-mesh codend were significantly lower than those estimated for the escape windows and sorting grid codends: almost 62% of cod and 79% of haddock escaping from the diamond-mesh codend did so at depth, whereas approximately 75% of cod and 93% of haddock escaped from the other two systems at depth. The partitioned selectivity parameters, l₅₀ and SR, were estimated and modelled for each of the three phases of the towing process and used to compare the selective properties of the three selection systems. A comparison was also made with a traditional whole haul selectivity analysis. Finally, bulk catch was added to the model as an explanatory variable and its effect upon selectivity was measured in both, partitioned and whole haul estimates.     

基于双重拔靴法拟合拖网网囊选择性的不确定性

研究拖网网囊选择性对于渔业的科学管理和资源的合理利用都有重要意义。拖网网囊选择性的研究一般先开展捕捞实验,然后使用选择性模型拟合和处理实验数据,最后得出选择性参数和选择性曲线。主要的选择性参数包括50%选择体长(L50)和选择范围(SR)。但是,仅仅提供选择性参数和曲线是不够的,还需要拟合它们的不确定性,并以置信区间的形式展现。双重拔靴法是拟合网囊选择性不确定性的重要方法。本研究在阐述双重拔靴法基本原理的基础上,选取2组实验数据为实例,用常规方法和双重拔靴法进行拟合,通过比较拟合结果诠释双重拔靴法的特点和优势。结果表明,双重拔靴法能够很好地拟合网囊选择性参数L50、SR和曲线的不确定性。同时,本研究还就应用双重拔靴法的重大误区进行了分析和讨论。本研究可为研究人员开展渔具选择性研究提供方法参考。     

Measuring the height of the fishing line and its effect on shrimp catch and bycatch in an ocean shrimp (Pandalus jordani) trawl

The relationship between fishing line height (FLH), shrimp catch and bycatch in a semi-pelagic ocean shrimp (Pandalus jordani) trawl was investigated using a newly developed recording inclinometer. The inclinometer was effective at measuring FLH and indicating trawl performance deficits. FLH was determined to be stable during a haul and also between hauls within a given footrope and groundline configuration. FLH was readily adjusted with simple modifications to the footrope “dropper” chains. Inclinometer data showed that FLH can be unequal between double-rigged nets of identical configuration. Shrimp catch and the bycatch of flatfish and juvenile rockfish varied inversely with FLH, suggesting FLH can be adjusted to equalize the catch of shrimp, flatfish and juvenile rockfish between two double-rigged shrimp nets.     

Bathymetric distribution of demersal fish in the Aegean and Ionian Seas based on generalized additive modeling

Good knowledge of the spatial distribution of fish is critical to stock assessment and successful fisheries management. Depth is often the main gradient along which faunal changes occur when analyzing shelf and upper slope assemblages, and thus knowledge of the bathymetric distribution of fish species is of great importance. The depth distribution of 16 fish species (Chelidonichthys lucernus, Helicolenus dactylopterus, Hoplostethus mediterraneus, Lepidorhombus boscii, Lophius budegassa, Merlangius merlangus, Merluccius merluccius, Micromesistius poutassou, Mullus barbatus, Mullus surmuletus, Pagellus erythrinus, Peristedion cataphractum, Phycis blennoides, Serranus cabrilla, Trigla lyra, and Trisopterus minutus) of the Aegean and the Ionian Seas was evaluated by analyzing experimental bottom trawl data, using generalized additive modeling (GAM) techniques. A variety of bathymetric distribution patterns was observed. The main bathymetric zone of each species was defined based on the modeled relative density. Specifically, the lower and upper limits of the main bathymetric zone of each species were defined as the depths where the estimated relative density of the species becomes less than 1% of the maximum. This definition is proposed as a better and more informative alternative than reporting the minimum and maximum depths of encounter.     

Tailoring codend mesh size to improve the size selectivity of undifferentiated trawl species

Tiger flathead, Neoplatycephalus richardsoni (Castelnau), and sand flathead, Platycephalus bassensis Cuvier, are undifferentiated and managed with a common legal minimum length (LML). The Commonwealth Trawl Sector (CTS) and the Tasmanian Danish‐seine fishery (TDSF) use a minimum codend mesh sizes of 90 and 70 mm, respectively. The codend mesh size should be tailored to the LML, which is based on the length of first maturity of females (M₅₀). This study found the length–girth relationship of N. richardsoni and P. bassensis was not significantly different. Using the covered codend method, these two species had 50% retention lengths (L₅₀) of 294 ± 2 and 307 ± 3 mm, in 70‐mm and 90‐mm codends, respectively. L₅₀ estimates for mesh sizes from this study and others produced a curvilinear relationship: y = 120 ln(x) ? 214, with an r² of 0.8504. The size at maturity (M₅₀) for female N. richardsoni was 337 mm, which is larger than the estimate for female P. bassensis (247 mm). There is a mismatch between the estimates of L_50, the estimates of M₅₀ and the LMLs in each fishery, leading to suboptimal exploitation of female Platycephalus. The model produced in this paper recommends a codend mesh size of 98 mm for both fisheries to exploit Platycephalus species sustainably.     

Capture efficiency of a multi-species survey trawl for Snow Crab (Chionoecetes opilio) in the Newfoundland region

An experiment was carried out during September of 2001 to estimate the relative catching efficiency of the NAFC's (Northwest Atlantic Fisheries Centre) standard survey bottom trawl, the Campelen 1800 shrimp trawl, for Snow Crab (Chionoecetes opilio). Secondary trawls were mounted underneath the main trawl directly behind the footgear to estimate escapement of crab passing underneath the main trawl. Overall, the capture efficiency of the survey trawl was found to be much lower than 1.0 and highly variable, with variability being strongly related to substrate type and Snow Crab size. The probability of capture monotonically increased with crab size, and was higher on soft than on hard substrates. Three substrate-specific capture efficiency functions were defined; (1) efficiency was highest (about 0.73–0.95), and least dependent on crab size, on the softest mud substrate; (2) efficiency was lower (about 0.39–0.70) and linearly related to crab size throughout a soft mud–sand substrate; and (3) efficiency was lowest (about 0.05–0.33) on a variety of hard substrates within relatively shallow water. Our models over-estimated true efficiency because secondary trawls were not fully efficient, i.e., crab escaped underneath their footgear. Nevertheless, our experiment demonstrated that substrate type can affect capture efficiency and annual changes in trawl configuration or in Snow Crab distribution in relation to substrate type will affect survey biomass estimates. Additional studies using various approaches could provide a basis for standardizing survey catches for varying capture efficiency effects and refining survey-based estimates of biomass and exploitation rates of Snow Crab.     

Evolving and devolving bycatch reduction devices in an Australian penaeid-trawl fishery

Since bycatch reduction devices (BRDs) were first tested in Australian penaeid-trawl fisheries almost 20 years ago, there have been ongoing refinements to a discrete range of legislated designs; mostly in response to dynamic fishery-specific operational requirements. This study assesses the relative utility of the latest versions of two generic BRDs that have evolved/devolved in an estuarine fishery. Compared with a codend containing no BRD, a simplified Nordmøre-grid and radial escape section significantly reduced the total weight of bycatch by 74 and 36%, respectively. Catches of the targeted school prawns (Metapenaeus macleayi) were also reduced by 4 and 11%, respectively, although, at least for the Nordmøre-grid, these losses were apparently offset by the practicality of the revised design in terms of a simpler and cheaper construction. The presence of large quantities of displaced submerged plants (e.g. Egeria densa and Hydrilla verticillata) on the trawl grounds at certain times (which preclude using the Nordmøre-grid) may similarly compensate for the reduced efficiency of the radial escape section. In addition to quantifying the performance of the revised BRDs, the results provide direction for their future refinement in this and other similar fisheries.     

绳拖网网口结构设计和装配工艺

本文主要阐述适用于我国东海和黄海渔场的双船两片式底层绳拖网网口结构设计，上、下纲和悬纲长度计算方法及其装配工艺。根据绳拖网的上纲和下纲的工况特点，上纲呈抛物线状；下纲呈悬链线状，分别依据其各自的曲线方程组求算。关于悬纲和上、下纲之间的装配工艺，是按我国底层拖网特点设计的。经生产实践，证实这些设计和装配都是合适的。     

The fishery for Nephrops norvegicus (Linnaeus, 1758) in the central Adriatic Sea (Italy): Preliminary observations comparing bottom trawl and baited creels

The Norway lobster, Nephrops norvegicus, is mainly targeted by bottom trawlers whose selectivity (mesh size = 40 mm stretch) with respect to this species is poor producing large quantities of discards both in terms of demersal fish species and undersized Nephrops. The Pomo pit (central Adriatic Sea) is a very important nursery area for European hake, Merluccius merluccius, but is also the main trawling ground for the Adriatic fleet exploiting Nephrops. The use of baited creels may constitute an alternative to bottom trawling. This paper summarises experimental data collected from the western Pomo pit in order to compare bottom trawl and baited creel catches with the aim of evaluating the ecological consequences of creeling. Because of high scavenger activity in the area, the effect of this on the creel fishery was also investigated.A large fraction of the trawl catch was composed of juvenile undersized individuals of commercial species, the same was not true for creels. Composition of Nephrops catches differed markedly between creels and the bottom trawl. Scavenger activity was very high. The results indicate that ecologically the use of baited creels appears to be a valid alternative to bottom trawling but, unfortunately, may not comprise an economically viable solution in the investigated area, which is 30 nautical miles offshore.     

Illumination of trawl gear by mechanically stimulated bioluminescence

The role of bioluminescence stimulated by moving fishing gear was investigated using an ISIT video camera mounted inside a pelagic and demersal trawl. Data presented here show that trawling is responsible for generating bioluminescent light. In two Norwegian fjords, the mean number of bioluminescent events per metre of head rope per second were 5.14 ± 2.17 and 2.39 ± 1.0 for a pelagic trawl at ∼270 m depth and a demersal trawl at ∼500 m depth, respectively, travelling at 2–4 knots. By extrapolation of these data, thousands of point-source flashes occur over the entire mouth of a trawl every second during a tow, considerably increasing its visual presence to fish. The occurrence of stimulated bioluminescent flashes on a trawl in deep-water may provide sufficient illumination to permit herding in the absence of ambient space light. The aim of this study is to bring to the attention these preliminary observations of mechanically stimulated bioluminescence by mobile trawl gear as possible visual detection stimulus for fish and its potential to affect catch rates by altering herding, escapement and avoidance behaviour.     

广东大鹏湾虾拖网网囊网目选择性研究

虾拖网渔业是南海海域重要的作业方式之一,但尚未有关该渔业网目尺寸选择性的报道。为评价虾拖网渔具的渔获性能、合理利用虾类资源,在广东大鹏湾进行了网目选择性试验。试验采用双支架虾拖网作为试验网具,并使用套网法估算了网目内径为35、40和45 mm的网囊对主捕虾类的选择性。结果显示,虾类渔获中刀额新对虾、须赤虾、周氏新对虾和日本囊对虾渔获数量较多,比较以往资料,刀额新对虾和周氏新对虾优势体长、平均体长均有所下降,其它2种虾类平均体长未见明显变化;选择性分析结果显示,主捕虾类的L50随网目尺寸线性增大。根据试验结果并结合历史资料,笔者认为现阶段35 mm的网囊网目尺寸对于保护上述4种虾类资源是比较合适的。     

Effects of codend circumference and twine diameter on selection in south-eastern Australian fish trawls

Two experiments, done in a south-eastern Australian trawl fishery targeting school whiting (Sillago flindersi: Sillaginidae), examined the relative efficiencies and selectivities of five codends and extension sections made from double-twine, 90-mm (inside stretched length) mesh netting. All extension sections were made from 3-mm diameter twine and were 100 meshes long and 100 meshes in circumference, while the codends were 25 meshes in length. The first experiment tested three codends made from 4-mm diameter twine: one with a circumference of 100 meshes and two of 200-mesh circumference, with one of the latter incorporating two cross-sectional joins in its extension piece. The second experiment compared two 200-mesh circumference codends, one constructed from 3-mm diameter twine and the other from 5-mm diameter twine. The codends were alternately fished with a small-meshed control. The results showed a general trend of reduced selection by the 200-mesh circumference and thicker twined codends, and especially by the industry-preferred 200-mesh circumference codend constructed from 5-mm diameter twine. Experiment 1 found that the 100-mesh codend caught significantly fewer school whiting, retained catch and total catch than did the two 200-mesh codends, and the 200-mesh codend with the modified extension section caught significantly fewer school whiting and retained catch than did the 200-mesh codend with the straight extension. In the second experiment, the 200-mesh 5-mm twine codend caught significantly more total and retained catch, school whiting, and longspine flathead (Platycephalus longispinis: Platycephalidae) than did the 200-mesh 3-mm twine codend. Across all codends, the smallest lengths at 50% probability of retention (L₅₀) were estimated for longspine flathead, redfish (Centroberyx affinis: Berycidae) and longfin gurnard (Lepidotrigla argus: Triglidae) in the 5-mm 200-mesh codend. The limited size range of school whiting resulted in unreliable estimation of selectivity in the 200-mesh 5-mm twine codend, but this design was the only one that retained a substantial proportion of individuals below the minimum marketable length (15 cm fork length). While the 200-mesh 5-mm twine codend retained commercial quantities of school whiting, it seems far from optimal. It is suggested that a more efficient design comprising possibly smaller, square-shaped meshes should be developed and used in conjunction with temporal, spatial and catch restrictions.