Methods

The research spanned two years. In the first year's study, the compound grate tested was mounted in the net upside down so that the small bar spacing was at the top of the grate and the larger bar spacing was at the bottom of the net with the escape hole in the bottom. The upper edge of the cod end is attached to the grate at the juncture between the small bar spaces and the larger bar spaces. All shrimp and finfish were directed to the top of the grate where the small shrimp and finfish were able to pass through the small bar spacing (7/16") and exit the net. The larger fish and marketable shrimp worked their way down to the larger bar spacing (3/4") where the shrimp and some fish would pass through into the cod end and lobsters and remaining fish worked their way down and out through the escape hole. Two different grate surface areas of small bar spacing were tested. One of the compound grates had one quarter of the grate small bar spacing (10" of the 40" grate were spaced 7/16" apart) while the other compound grate was comprised of half the grate with small bar spacing (20" of the 40" grate were spaced 7/16" apart). A second cod end of smaller mesh was constructed to fit behind the small bar space panel to collect what was escaping through these bars on over the escape hole to determine if shrimp were being lost out through this hole.

The compound grate was tested using a trouser trawl, a dual cod end trawl split vertically down the middle (Figure 1). One lengthener/cod end held the experimental device while the other cod end was used as the control in a series of experiments. As there is always a question concerning bias in a split trawl as to weather the two halves fish equally, both lengtheners and both cod ends were attached to the trawl with a quick release zipper so that they could be switched between sides.

Each of the panel sizes (1/4 small bar spaces and 1/2 small bar spaces) was tested by a series of tows against a standard Nordmore grate and each side had a standard 1 3/4" diamond mesh cod end. Each panel size was also towed with and without the small mesh cod end behind the small bar space panel. The species weights and length frequencies in the cod ends were compared to determine any differences in the catch. The length frequency in the regular cod end was also compared to the length frequency in the small mesh cod end to test against the possibility of back pressure affecting the flow through the bars of the small bar space panel. The tow series to conduct these tests switched the standard and experimental units between sides to reduce any possible side effect with the trouser trawl (Table 1).

In the second series of tests the next year the study built upon the results of the first year and modified the grate structure to improve loss of finfish and small shrimp and retention of large shrimp. We used the half grate configuration, elongated the grate and tipped the upper, small bar space half about 10 degrees more towards the front of the net to improve movement of shrimp and fish down that portion of the grate. With this change, two bar space widths were tested in the upper, small bar space section, a straight 7/16" grid and a trapezoidal grid with tapered bar spacing increasing from 5/16" to 1/2" front to back. Each of these bar space configurations were tested with diamond mesh in both the lengthener around the grate and the cod end, with square mesh in the cod end only, with square mesh in the lengthener only and with square mesh in both the lengthener and cod end to see which combination resulted in the best separation of fish and small shrimp from the market shrimp. The extra sets of tow with the small mesh bag behind the small bar space grate and with small mesh bags behind the small bar space grate and over the escape hole were not conducted in the second year due to time and funding constraints. The tow series to conduct the second years' tests switched the standard and experimental units between sides to reduce any possible side effect with the trouser trawl (Table 2).

In looking at the boat procedure there were a couple of things that had to be determined. The first was the total time to tow the net. The average tow time for the Maine shrimp fishery from 2001 through 2004 was 2.1 hours per tow with fishermen getting in about 3 or 4 tows a day. The other factor in determining the duration of the tow was to get as many replications as possible with the budget and time we had. One hour tows were chosen in order to remain as close as possible to realistic fishing conditions while getting in 3 or 4 extra replications per day. The tow time started at the started. The date, starting time, ending time, starting latitude, ending latitude, starting longitude, ending longitude, starting depth, ending depth, tow speed, wire out and gear type on the starboard and port sides of the trouser trawl were all recorded for each tow. Once the cod ends came on board the boat the starboard cod end was emptied onto the sorting table. The port cod end was emptied into fish trays. The shrimp were then separated from the fish in the starboard catch. The fish were separated further by species and a total weight by species was recorded along with the individual lengths of each fish. In the cases where there were quite a few fish a weighed sub-sample was taken. The total shrimp catch was then weighed and a 1-kilogram sub-sample of shrimp was brought back to the lab for further analysis. Once done with the starboard side the process was repeated for the port side catch. The catch from the mesh bags behind the small bar space grate, or over the escape hole, when employed, was sorted and measured in a similar fashion.

During the lab analysis of the 1-kilogram sub-sample the shrimp are sorted by species and the Pandalus borealis are sorted by sexual stage. For the non Pandalus borealis species total weights for each species are obtained along with individual carapace lengths. For Pandalus borealis the total weights for each sexual stage is recorded. Then for each sexual stage the individual carapace lengths are recorded. This information is used to generate sex-specific length frequency distributions, thus providing information on the size and age composition of each of the catches.