A stretch of the River Drwca, near the Czarci Jar farm8, was chosen as the site for the tests (Figure 1). The characteristics of the stretch were as follows:
8 Polish Angling Association, Czarci Jar, Poland
| - | Length | 100 m |
| - | Width | 2–4 m |
| - | Depth | 0–9 m |
| - | Conductivity | 333 uS/cm |
| - | Temperature | 7°C |
| - | Flow speed | 0.18 m/s |
| - | Bottom | sand and mud |
| - | Emergent vegetation: | Acorus calamus, providing shelter for the fish. |
The experimental stretch was isolated by traps at both ends (Figure 2), and all the native fish were removed electrically before the tests: these comprised rainbow trout which had escaped from the fish farm, tench, carp and pike.
Two species were used:
- Rainbow trout (Salmo gairdneri Richardson)
- Common eel (Anguilla anguilla L.)
These two species were used in numbers and sizes appropriate to a trout stream. The fish had been marked with plastic tags 5 × 10 mm attached anteriorly to the pectoral fins (Figure 3). Each tag was numbered according to the batch used for a test. Three colours were used for the tags to distinguish size classes.
The composition of each batch of fish is shown in Table I.
The batches of fish had been prepared and held before the start of the tests at the fish farm at Czarci Jar. A reserve supply of fish had been arranged for every batch in case of sickness or death occurring before the tests. Every care had been taken over aerating the travelling tanks and expediting their transport in order that the fish might reach the experimental sector in perfect condition.
The machines intended to be tested and the one used for clearance have been respectively designated by the letters A, B, C, D and E; their characteristics and performance are shown in Table II. The interelectrode resistances were measured in the condition of use by the relationship I/R, when coupled to the alternator E. The diagrams showing pulse shapes of the various machines were produced subsequently from oscillograph traces supplied by each expert with his apparatus working into the interelectrode resistance determined on the site (Figure 4).
Certain machines possessed peculiarities which are worth mentioning.
Machine B
The cathode makes contact with the water at three positions; two of these are formed by a nonslip chain overboots worn by the fisherman over his waders, and the third by a copper strip dragged behind the fisherman; the area of this one can be reduced according to the conductivity of the water.
Machine C
The cathode consisted of three linked bamboo rods (each 1 m long) floating bare wires which hung down in the water. The wires were each about 30 cm and there were about 13 on each bamboo. This was intended to form a barrier across the stream behind the fisherman, and was dragged by two assistants, one on either side of the stream.
Machine D
The cathode consisted of a copper tube about 30 cm long and 4 cm in diameter pushed onto the end of a staff. This type of cathode is meant to be used as a wading staff and can be used also to drive fish out of shelter.
Machine E
An electronic safety device automatically reduces the voltage from the working value of 220–280 volts to 12 volts - the safety voltage when the anode is out of the water; when the anode is again immersed, the working voltage is re-established. This obviates accidents should the anode touch a fisherman.
These are shown diagramatically in Figure 5.
Machines A and E
The generator remains on the bank; the cathode is placed in the stream near the generator. The anode is connected by a cable of adequate length. In this technique, four operators were needed, one of whom handled the cable.
Machines B and D
These are carried on the back of the user, and only three operators are necessary.
Machine C
The generator is carried by one operator who also handles one end of the cathode “raft”, the other being held by a second man on the opposite bank. Five operators are necessary.
For each test the operations were as follows:
- Decanting the batch of fish and distributing them throughout the test sector
- Waiting ten minutes to allow the fish to settle down and take shelter
- Fishing with the machine under the test, which took 20 minutes; the sector was worked from the lower to the upper end, in order to keep the water clear for the fisherman
- Lifting the traps at both ends of the sector
- Clearing the sector by fishing with machine E
These operations occupied 1 hour 30 minutes.
The fish caught in the course of each test were immediately carried to a bench where a special team identified and counted them.
The tests took place in the order and at the times and dates shown below:
| Date | Starting time for test (h) | Test number | Machine |
| 1.10.69 | 12.40 | 1 | A |
| 14.03 | 2 | A | |
| 15.39 | 3 | B | |
| 2.10.69 | 8.50 | 4 | C |
| 10.14 | 5 | D | |
| 11.40 | 6 | B | |
| 13.10 | 7 | A | |
| 14.35 | 8 | E |
Test No. 1 was considered as a “blank”, and used to determine the fishing time. Tests 1, 2 and 7 on the one hand, and 3 and 6 on the other, carried out with the same machines, allowed the variation in the results for a single machine to be studied. Finally, test 8 concerns the machine used for clearing the channel of fish; it was meant to compare the results with those obtained with other machines.
All the tests, except the sixth, were carried out by the same team of technicians, who were familiar with practical electrical fishing from long experience.
Several factors conspired to disturb the course of operations:
- an excessive stocking with fish from the reserve in the course of one of the first three trials; it appears from the analysis that the results of Test 2 are affected;
- an escape of fish from the sector past the traps, which had been undermined by the water flow; analysis of the results suggest that these escapes were particularly large during Test 7;
- a breakdown owing to the contact failure in the case of machine B during Test 6, at the upper end of the sector, at the time when the trout should have been concentrated between the trap and the fishermen;
- the intermittent rain and occasional lightning which to some extent hampered the fishermen during some of the tests.
Originally, it had been intended to measure fishing efficiency only in the case of fish which had not been electrically fished previously. However, as the clearance fishings were incomplete, fish belonging to earlier batches accumulated in the stream, and this had permitted two categories of efficiency to be determined for each machine:
- The efficiency which we designate “A” (“At” for trout, “Ae” for eel) which applies to fish liberated for the particular test;
- The “B” efficiency (“Bt” and “Be”), applying to fish remaining in the river after having escaped both the machine under test and the subsequent clearance. It, therefore, represents the efficiency of the third fishing.
Reference to Tables III and IV, containing all the results, will show these efficiencies calculated thus:
Efficiency A9
| For test 1 | (Batch 1, column 6 total) × 100 (Batch 1, column 4 total) |
| For test 2 | (Batch 2, column 11 total) × 100 (Batch 2, column 4 total) |
| and similarly for succeeding tests. | |
Efficiency B9
| Inapplicable to test 1 | |
| For test 2 | (Batch 1, column 11 total) × 100 (Batch 1, column 4 total) - (Batch 1, column 10 total) |
| For test 3 | (Batch 2, column 16 total) × 100 (Batch 2, column 4 total) - (Batch 2, column 15 total) |
etc.
This is shown in Table III for trout and Table IV for eels.
Table V shows the At and Bt efficiencies for each test for each machine. The tests are arranged in order of decreasing At efficiency. Column 7 shows the numbers of trout recaptured during all the tests, the clearances, and found in traps, divided according to batch.
Table VI shows the At efficiencies for each test for each machine. Bt efficiencies by size classes are not significant owing to the small samples, and have not been shown in Table VI.
Table VII shows the Ae and Be efficiencies for each test for each machine. The order in the case of the Ae efficiencies is the same as that for trout.
Table VIII shows the Ae efficiencies for each test for each machine. As in the case of trout, the Be efficiencies are not shown owing to lack of data.
In order to obtain a correct interpretation of the differences in efficiency between the various machines, it would be necessary to make a statistical comparison of the variability of the results, both for each machine and between the various machines. Unfortunately, the factors mentioned in paragraph 2.21 introduced into the experimental data uncertainties which preclude a statistical interpretation. We must, therefore, restrict ourselves merely to commenting on the results, accepting that they have only a doubtful validity.
Examination of Tables V and VI reveals considerable differences in efficiency between the machines. These differences could be the results of simple chance or artifacts, but possibly they may appear to be logically distributed.
Taking Table V first of all: the results have been arranged in order of At decreasing efficiency values, by test. The order, taking machine by machine, remains the same, which goes to show that the variation in the results between tests for a given machine is insufficient to affect the order of placing. This is not rigorously repeated in the case of the Bt efficiencies, though an analogy might have been expected that a machine showing a higher At efficiency would probably display a higher Bt efficiency also. Now, certain of these Bt efficiencies appear to be dubious; these are underlined in Table V. The reason for the dubiety is as follows.
The value 11.72 (column 5) seems to be underestimated for machine E, which was chosen for making the clearances on account of its relatively great and constant efficiency (32.16 percent on average, calculated from Table III clearances). It is now acting with an efficiency Bt on the fish remaining in the stream after the previous test, namely, test 7. Referring to column 7, it is seen that of this batch of trout, only 67 have been recaptured by all efforts (2 tests, 2 clearances and the traps) out of 160 trout put in 10. This is by far the smallest proportion. As it has been established that trout succeeded in passing under the traps, it may be questioned whether the percentage escaping was not exceptionally high in the case of this batch. This would therefore result in a slight underestimate for the efficiency in test 8, which would have been calculated on a theoretical number of fish greater than the true number. If this hypothesis is correct, the At efficiency of test 7 would also be underestimated, diminishing the variations in the results for machine A.
The value of 17.31 percent in column 5 appears to be an overestimate. It refers to a Bt efficiency calculated on the fish of batch 2. Now, column 7 shows that for this test 140 out of 160 fish were recaptured, a proportion very great compared with the others10. During the proceedings (para. 2.21) it was reported that an excessive number of trout was introduced in error in one of the three first tests; this probably occurred in test 2, and the Bt efficiency for test 3 would then have been an overestimate. An excessive stocking for test 2 would consequently have resulted in an overestimate of the At efficiency for this test, and consequently a reduction in the variations between the At efficiency values for machine A.
There is a surprisingly large difference between the At efficiency of 45 percent and 30 percent for tests 3 and 6 carried out with the same machine (B). It has already been stated in the description of the course of the tests that this machine was out of action for several minutes when the electrodes were near the upstream trap, and a considerable number of large trout was collected in the space between During the repair, these trout could have escaped downstream, which would explain why the efficiency was particularly low for large trout (columns 5 and 7, Table VI).
Thus, the anomalies appearing in Tables V and VI appear to be due to known artefacts rather than variations caused by the experimental method itself.
There are additional arguments lending support to the results. Each size category can be considered as an independent batch, enabling each test to be considered to some extent on a triple basis. For the At efficiencies, it is obvious that the order shown for the machines in Table VI according to fish length remains the same as that for the fish irrespective of length. It is also obvious that for each machine the At efficiencies increase with fish size, as has been well known for a long time. In the case of Bt efficiencies by size classes, the very small number of specimens robs the results of any significance.
The tests are listed in the same order as for trout: that of decreasing At efficiencies. The order for the machines remains the same for eels as for trout as far as the Ae efficiencies are concerned. If, however, the results are examined in detail, they appear far less homogeneous than those for trout. In particular, the order of Ae efficiencies by size classes is not the same.
The efficiencies for each test do not increase with length, contrary to the case with trout. This may be explained by the range of the size classes, 5–30 cm for trout, only 20–50 cm for eels, but there may be other reasons concerned with the physiology of the animals.
The great disparity of the results for eels appears to depend, among other factors, on the relatively small number of eels in each test: 60 eels as against 160 trout.
The traps were intended not only to prevent the escape of fish from the sector fished, but also to show the driving effect produced by any particular machine. They were only partly successful, as trout were able to pass under them.
Table IX shows that the proportion of fish taken by the traps is small; the results have no great significance.
It can be seen, however, that the proportion of fish taken in the downstream trap is always much greater than in the upstream trap; it is possible that the downstream captures are due to drifting fish shocked during the fishing operation.
The method tested has shown that a comparison between the efficiencies of fishing machines could be made, and apparently consistent results obtained. However, variations in the results for a single machine, probably due to the experimental artefacts mentioned above, prevent us accepting the conclusions as definite.
The tests demonstrated that fish become progressively harder to catch; the following figures show the great reduction in efficiency between the first and third fishing:
| Mean Efficiency A: (first fishing) | At Ae | = = | 27.50% 25.04% |
| Mean Efficiency B: | Bt Be | = = | 8.11% 11.36% |
In the case of machine E, there are two successive fishings: the test fishing and the clearance, which constitutes a second fishing. These provide values of:
- 65.63% for the first, and
- 32.73% for the second fishing.
This diminution of efficiency in the course of successive fishings is a serious factor in estimating the efficiency of a machine or an electrical regime. In fact, it is likely that this diminution varies between various electrical regimes and fishing methods. In fish population estimations, which are the chief use of electrical fishing, it is important to maintain a high efficiency throughout several fishings; in fact, the de Lury method is based on the constancy of fishing efforts over several successive fishings.
Lastly, these tests have shown that the efficiency also diminished with diminishing size among the trout; this result was expected, and it did not seem worth while to complicate the organization with additional length categories. This does not, however, seem to be the case with eels. These are more easily stunned than trout, and thus less able to approach the anode; this phenomenon is correspondingly more marked as size increases, and it may be that efficiency is greater in the case of small eels. This must be taken into consideration at the start when selecting the correct size of eels for the test batch.
Another criticism which may be made of the method used was the low efficiency of the clearance operations (32.86 percent on average) compared with anticipation. The probable cause is insufficient smoothing of the current generated by the machine used. This resulted in leaving fish in the river in increasing numbers as the tests went on, and thus created circumstances more and more unfavourable for successive machines. But even if the clearance machine had produced pure direct current, its efficiency would have been inadequate, as the clearance operation represented a second fishing at correspondingly reduced efficiency. For this method to be rigorously applied, it thus appears necessary that the channel be drained and totally cleared of fish after each test. These points are taken up again in the succeeding paragraph.
We have just seen that the channel should be drainable. It should in addition include natural features: shelters, rocks, weeds, water speed, etc. If the channel is too easy to fish, there will be insufficient differences between various tests.
The use of traps, both upstream and downstream, appears to be worth retaining; the lower one collects shocked fish washed downstream, and the upper one ensures that fish which have escaped upstream in front of the anode shall not be too easily captured, as they would be if finally confined between the anode and a simple grid.
The composition in terms of species and sizes depends on the object which is intended. For example, if the efficiency of a current for capturing small trout is being studied, it is useless to release larger trout into the test section. In our view it is better to simplify the sampling by using fish of the same size, and to compensate for this restriction on the information by using a great number of fish. With regard to number, 200 fish per test appears to us a suitable number. The tests carried out with eels were not significant because their numbers (60) were inadequate, and also because the technique for catching eels is different from that for trout.
The study of the A efficiency must obviously be continued. It permits comparison between machines used for a single fishing (excluding population estimates), as for example, a salvage operation or commercial fishing, but it is insufficient to characterize a machine making population estimates. As we have already pointed out, catchability diminishes considerably from one fishing to the next, and this diminution may be greater or less according to the machine used. In general, it seems considerably more evident with pulsed currents than with continuous direct current. It therefore appears important to measure the B efficiency for a second fishing; in the tests carried out at Czarci Jar, the B efficiency related to the third fishing, after the clearance fishing.
The B efficiency could be measured in the following conditions:
- the fish caught in the first sweep should not be replaced in the water;
- the B efficiency would be calculated so:
Fish caught in the second fishing
Total stocked fish less fish caught in first fishing
This has the advantage of approximating the test to the de Lury method.
The incidents in the field trial such as faults during the test, hazards of weather, etc., make it essential that a machine should be tested several times. This is equally desirable for the statistical interpretation of the results. It seems that three tests per machine make an adequate number in view of the limited time available for this kind of operation.
We have seen that emptying the channel would be an essential operation between each test and the next, but before undertaking a new test, it would seem necessary to make equated inventory of the fish caught:
- fish stocked - fish caught by the machine + fish collected during emptying.
If, in spite of the precautions taken, fish still remain in the channel after emptying - and this will probably occur with eels - it will be necessary to take account of them in the succeeding test. For this reason, it would be better to mark the fish intended for each test by fin clipping.
This rational method requires experimental arrangements which have been thoroughly proved. If the tests carried out at Czarci Jar did not provide definite results, it was partly because the experimental arrangements had not been tested in advance.
Moreover, it could only be carried out at a fish farm, where an artificial channel capable of being emptied, and a sufficient number of fish would be found. The possibilities of its use are therefore restricted.
We will conclude by repeating that in the absence of a fish farm it is always possible to use on a river and in natural conditions the method proposed by P. Lamarque. However, it is important in this case to be sure that the number of fish in the test river is sufficiently important to get significance.
