The method of calculation for the spread of the trawls, that is, the width of the mouth opening from the end of one wing to the end of the other, was as follows. With both the Young Fish Trawl and the Commercial Vendace Trawl the sweep lines were 15 m and the bridles were also 15 m, but the length of netting from the end of the wing to the cod end was 20 m in the Y.F.T. and 35 m in the C.V.T. Therefore for the Y.F.T. we have:
| Length of warp | 100 m |
| Length of sweep line | 15 |
| Length of bridle | 15 |
| Length of netting | 20 |
| Total | 150 m |
With 100 m of warp the distance between the boats was 60 m, so by proportion if the distance across the mouth of the net was x,
With the other lengths of towing warp (and corresponding distances between the boats given in Tables 1 to 3) we have:
| Trawl | Length of towing warp | Spread of the trawl | |
| Y.F.T. | 150 m | 7.5 m | |
| 125 m | 8.6 m | ||
| 100 m | 8.0 m | ||
| 75 m | 9.6 m | ||
| C.V.T. | 150 m | 12.2 m | |
| 100 m | 12.7 m | ||
The speeds at which the trawls were towed were determined primarily by accurate measurements of distance covered in a given time using a log and stop watch calibrated with throttle position on each boat. These calibrations were carried out during the preliminary day-time trials and gave 3.2 kn for the Y.F.T. and 2.7 kn for the C.V.T. These speeds, together with the duration of the hauls during the exercise, give distances which were clearly reasonably accurate when viewed in conjunction with the space available for trawling.
Previous work by the second author has shown that at the speeds given above, the distance between the head line and foot rope is 2.5 m in the Y.F.T. and 3.5 m in the C.V.T.
The figures given above in this section can be used to calculate the volume of water filtered given in Table 4.
Table 4
The volume of water (m3) filtered per minute
| Warp length | Y.F.T. | C.V.T. |
| 75 m | 2 371 | - |
| 100 m | 1 976 | 4 391 |
| 125 m | 2 124 | - |
| 150 m | 1 852 | 4 218 |
The density of fish per 1 000 m3 has been calculated from Table 4 and the number of fish caught in each haul. This has been added for each haul to Tables 1, 2 and 3.
The variability in catches from replicate trawl hauls is always large. From the data collected in the EIFAC experiments in 1976, Bagenal (1980) calculated that the 95 percent confidence limits expected for the mean of 5 trawl hauls would be the mean catch multiplied and divided by 2.7. The catch per 1 000 m3 in Tables 1, 2 and 3 are based on single hauls; the confidence limits for these would be very large indeed. In comparing trawl-based estimates with those that are acoustically determined, the variability of the former must be borne in mind. This variability is primarily due to the patchiness of the fish distribution which is particularly marked in shoaling species.
