Previous Page Table of Contents Next Page



Christopher Moriarty
Department of Agriculture and Fisheries, Dublin 2


The operation of trains of small fyke nets (diameter of opening 50 cm) for the capture of immature eels in lakes, rivers and estuaries is described. The nets have proved to be efficient gear for collecting samples of eels of lengths greater than 40 cm while sufficient specimens of lengths 30 to 40 cm are retained to allow useful conclusions to be drawn. Relatively small numbers of other fish species are captured.

The length of the leader should be greater than 6 m in still water, though shorter lengths down to 3.3 m are effective. A trap of length 1.8 m was found to be adequate and longer traps appeared to offer no advantage. Variations in local populations, in preference of depth and in activity between large and small eels, require that the exact method of fishing must be controlled to allow comparisons to be made between different populations.

Examples of the interpretation of figures are given. Information on relative population densities, age of maturing and size at maturity, intensity of exploitation and feeding regimes of eels may be obtained.


On déorit la mise en opération de nasses (verveux multiples) (diamètre d'ouverture 50 cm) pour la capture d'anguilles immatures en lacs, fleuves et estuaires. Il s'est avéré que ces filets sont efficaces pour recueillir des échantillons d'anguilles de plus de 40 cm de longueur tout en retenant un nombre suffisant de spécimens de 30 à 40 cm, ce qui permet des conclusions utiles. Un nombre relativement faible d'autres espèces de poisson est capturé.

La longueur de l'aile devrait être de plus de 6 m en eau calme, même si des longueurs moindres, jusqu'à 3,3 m, sont encore efficaces. Une nasse de 1,8 m de long s'est démontrée adéquate, alors que des nasses plus longues n'ont pas donné de meilleurs résultats. Les variations chez les populations locales, en ce qui concerne les préférences de profondeur et l'activité entre les grandes et les petites anguilles, obligent au contrôle précis de la méthode de pêche pour permettre des comparaisons entre des populations différentes.

Des exemples d'interprétation des chiffres sont illustrés. On peut en retirer des renseignements sur les densités relatives de populations, l'âge auquel la maturité est atteinte et la taille à la maturité, l'intensité d'exploitation et les régimes alimentaires de l'anguille.





3.1 Mesh Size
3.2 Length of Leader
3.3 Effect of Direction of Current
3.4 Catch-per-unit of Effort
3.5 Length Distributions
3.6 Age Distributions
3.7 Local Movements
3.8 Food
3.9 Other Fish Species




The Summer Fyke Net has been the principal instrument for the study of eel populations in Ireland since 1965. This paper describes the method of operation and gives examples of the results obtained and their interpretation. On account of their long life span and individualistic habits eels show enormous variations both within and between populations and little progress has been made in the application of standard methods of population assessment. It is believed, however, that much useful information on the eel stocks of a water- course may be gathered by the methods described here.


The summer fyke is so called because it is effective in capturing feeding eels in the warmer months of the year. Mature eels on migration are seldom caught when the net is totally submerged, the usual method of fishing. A single unit of the net consists of two fyke net traps connected by a leading net with the openings facing each other. A detailed description of the net is given by von Brandt (1971).

The nets are operated in long chains and professional fishermen in the Baltic may use as many as a thousand together. In the present experiments where, frequently, an assistant had to work on his or her own, trains of eight nets were found to be convenient. On average they yielded a catch of eels which could be examined in the course of a day's work. When larger teams of workers were available it was found that an average of eight or ten nets per man gave a satisfactory workload, permitting the detailed examination of up to one hundred specimens per day. Eight nets when wet weighed about 16 kg and, packed in a polythene bag, made a convenient bundle for one person.

In this paper the term “net” refers to a single pair of traps with their leader; the “trap” is therefore equivalent in catching power to half a net. The expression “net day” refers to the catch made by one net over a twenty-four hour period. The diameter of the opening of the trap, always slightly less than the height of the leader, was of the order of 50 cm. It is believed that the exact size of the opening has no effect on the catch.

In shallow still water the train of nets was attached at each end by about 2 m of light rope (25 mm circumference) to a sinker weighing 750 g. The sinkers were made by filling a 500 g food can with concrete in which a loop of wire was inserted for attachment of the rope. The position of one end of the net was marked by an inflatable buoy. In deep water a heavier rope was required because the entire train of nets, including mud or stones caught in them, may be suspended above the lake bed in the course of hauling. The heavy rope was desirable not only for strength but also because a light wet rope was painful to haul by hand under these circumstances.

In river water the upstream end of the train and the marker buoy were attached to a boat anchor weighing 2 kg. The downstream end of the train needed no more than the sinker. In tidal water both ends of the train were attached to anchors. Few or no eels were caught between high and low water marks and it was therefore necessary to set the nets at low tide to ensure that they were not left exposed.


3.1 Mesh size

The accepted size for codend mesh now appears to be 12 mm knot to knot. This allows the escape of a large proportion of unmarketable eels. Controlled experiments on the effect of mesh size were not made because it was not found possible to obtain a variety of meshes from the suppliers. The smallest mesh used was 5 mm and this net proved to be virtually ineffective. It caught far fewer eels than the wider-meshed patterns and failed to capture the small specimens which it had been expected to secure.

Unfortunately, a direct comparison between two different mesh sizes was possible on only one occasion. This was in northern Lough Corrib in 1968 and no significant difference was observed in the mean lengths of eels retained by meshes of 10 mm and 12 mm. In each case the smallest eel captured measured 28 cm. As will be seen from Table V, small eels were scarce in this region. It is possible that in a population which included many small eels an effect of this degree of difference in mesh size would be noticeable.

3.2 Length of leader

In this section it has been assumed that the length of the trap is immaterial. Von Brandt (1971) held that in a fyke net the tail of the eel should lose contact with the first funnel before the head enters the second, otherwise the eel may be alarmed and succeed in backing out of the trap. This may hold for migrating eels which appear to be more sensitive to obstructions. In the net which has been used for the last two seasons in this investigation, the distance from the end of the first to the end of the second was 73 cm. This net caught substantial numbers of eels of over 40 cm and smaller numbers of specimens of over 73 cm. A large proportion of Irish eels migrate at lengths of less than 73 cm so that small numbers of this size were to be expected. The greatest catch to date in one of these traps was 38 eels of which only one measured less than 40 cm. Five of them were more than 73 cm and therefore were in contact with the ends of both funnels. This, incidentally, was the smallest trap used in the course of the experiments. Eels of over 90 cm have been caught in these traps and larger specimens are extremely rare in Ireland. The overall length of this trap was 1.80 m and it is believed that there is no advantage in using a greater length.

Comparisons of nets of different leader lengths are shown in Table I. Significant differences either in the mean catch of eels per trap or in the relative numbers of traps with no eels and traps with one or more eels were observed in three cases. The first was in Lough Corrib in the first half of May, when 70 percent of traps with 9 m leader contained eels as compared with only 6 percent of thosewith leader of 6.5 m. At this time the water was still relatively cold and it was clear that eels were scarce either as a result of winter emigration from the area or because the majority had not begun their normal summer activity. When the whole season on Lough Corrib the same year was considered it was found that the mean catch per net of the shorter leaders was double that of the longer. The conclusion was drawn that under normal conditions the 9 m leader was unnecessarily long and offered no advantage; it possibly had a disadvantage in separating the nets by too great a distance. Possibly, the long leader acted more efficiently under conditions of low population or low temperature.

Also in Lough Corrib in the same season, a highly significant difference was detected between the catches of nets with leaders of 6.5 and 4.7 m. The longer leader resulted in a mean catch at least 50 percent higher than that of the shorter. On the other hand, in later experiments, no significant differences were observed between the catches made in nets of leader length 4.7 and 3.3 m either in still or flowing water. It appears, therefore, that in still water a leader of length 6.5 m is to be preferred to shorter ones but a length of 9 m is greater than required. Presumably the ideal length lies between 6.5 and 9 m. Controlled experiments on both leader and trap length would be very desirable but they would require a large and uniform body of water of a kind which is not to be found in Ireland.

3.3 Effect of direction of current

The traps which opened downstream in rivers caught approximately twice as many eels as those opening upstream. It is not known whether this difference was caused by a tendency for eels to move against the current when they met the leader or by the accumulation of debris which blocked the funnels of some of the upstream-facing nets. Under estuarine conditions the direction in which the traps faced made no difference to the catch. Figures for the River Blackwater are given in Table II.

3.4 Catch-per-unit of effort

Using gear of constant dimensions, comparison of the catch-per-unit of effort may be taken as a comparison of the population densities in different areas. Some precautions must be observed in view of the effects of time of year, depth and benthic conditions. Table III shows a selection of results from Lough Corrib in 1967 and 1968.

The catches in May and early June were consistently much lower than in the lower part of the summer, an observation made both on a local level (Dooros Bay) at different depths (1–4 and 4–10 m) and over a wider area. In some regions populations which were very much more dense than the mean were made; two such cases, Urkaunmore and Inishshanboe South are shown in the Table. Inishshanboe is an island approximately 500 m by 250 m and the remarkable discrepancy between an average catch to the east and more than double this to the south was recorded. In this case the most likely explanation lies in the nature of the terrain. The lake lies in the valley of a former glacier which moved toward the southeast. Material of a different nature from that on the south could have been deposited to the east of the island. The significance of the result in this case, however, is to indicate the danger of limiting the sampling area when a lake population is to be assessed. Short trains of nets set at random in many places are to be preferred to a single train in one place.

Taking the mean catch over the whole area in the late summers of 1967 and 1968, a close correlation was observed. This would suggest that, in spite of the considerable local variations, an effort of about 200 net days is adequate for a complete sample. In the two years in question, most of the work was done by a single operator and the sampling was spread out over a long period. Subsequently, teams of four or five assistants became available and the same effort could be achieved in four or five days. It has not yet been possible to collect extensive data on the effect of depth of water on populations but in general a decline in numbers has been noticed with increasing depth. In the interests of consistency, the greater part of sampling after 1968 has taken place in water between one and three metres. When an adequate sample has been collected in this way, experiments in deeper water have been made.

Comparative figures for a number of waters are shown in Table IV. In all cases nets of leader length 4.7 or 3.3 m, sometimes both together, were used. Positions around the lake shore were selected when possible by drawing a grid on a 1 : 10 560 scale map and selecting squares by the use of random numbers tables. However, weather conditions frequently necessitated the use of the nearest sheltered area to the base of operations.

The first group of waters: Lough Corrib, Lough Mask and Lough Carra, lay on the same river and the figures showed clearly the effect of lakes in impeding upstream migration so that smaller and smaller populations were found. Loughs Eonish and Key were rich lakes in the upper waters of their respective river systems and contained sparse eel populations.

The figures from the River Blackwater suggested that, where access to freshwater was unimpeded, eels showed a tendency to leave the brackish areas. At Careysville the populations sampled in the month of August in two successive seasons showed no significant difference in density. However, at Banteer some 55 km upstream, the density of population had been markedly reduced.

Finally the figures for the Wexford South Sloblands are given. The water here was slightly brackish and lay in an area of land reclaimed from the sea. It had no access to river systems and therefore eels were unable to leave it to travel upstream. It was also a very rich feeding area with large populations of Chironomidae and various small crustaceans. In 1970 it yielded a catch per net day of 15.6; by far the highest recorded in Ireland. The following year intensive commercial fishing took place resulting in the great decline in numbers indicated by the catch of 0.3 eels per net made in 1972. Commercial fishing was discontinued and sampling in 1973 showed a considerable recovery of the stocks, to 4.7 eels per net. The results of annual sampling will be used to advise the fishery owners when to begin operations again.

3.5 Length distributions

Eels of less than 30 cm were seldom retained by the standard mesh of 12 mm and those of less than 40 cm were not fully sampled. Length distributions are shown in Table V. In Lough Corrib a significant difference between mean lengths of samples taken in May and in July/August is shown.

The sample from Urkaunmore consisted of exceptionally large specimens and there were other isolated instances of gatherings of large or small eels. On the Corrib lakes, a gradual increase in mean length, caused by a decrease in the numbers of small eels, was shown with increasing distance from the sea. The South Basin was closest to the outlet of the lake with Lough Mask and Lough Carra further upstream on the same main river.

The South Sloblands figures collected before and after heavy commercial fishing showed clearly the decline in numbers of marketable eels. The Blackwater estuary lengths showed a more marked preponderance of small eels in May than that observed in the North Basin of Lough Corrib.

In the freshwater portion of the Blackwater, the expected decline in numbers of small eels with increasing distance from the sea was observed. An unexpected result, repeated in two seasons, was the marked drop in the proportion of large eels taken in successive nights of fishing. In August 1972, 51 percent of the catch were greater than 50 cm on the first night, compared with 16 percent on the second. In 1973, 44 percent of large eels were taken on the first night and only 23 percent two nights later. Possibly, the larger eels make more extensive foraging journeys and therefore are more likely to meet the net in the course of a single night.

3.6 Age distributions

Growth of eels in Ireland and other north temperate countries is frequently very slow and very irregular. Age determinations from otoliths can seldom be made with certainty and, ideally, very large samples of eels should be taken if precise results are required. In our experience it has not been possible to obtain such large samples, and, in most cases, collections of the order of one hundred specimens have been used. In examining age distributions it has been assumed that all of the eels in a given area of water belong to the same population. While the preference of eels of particular sizes for particular places has been shown, a tendency for individual eels to spend their entire lives in limited areas has not been observed. Fast-and slow-growing specimens have frequently been found in the same samples. It is therefore assumed that the population is randomly distributed on an age basis and in calculating growth rates the exact locality of capture and the time of year can be ignored.

Rates of growth seldom differed significantly between river systems. An exception was that of the Erne eels which grew faster than average. Their mean length at age ten was over 50 cm in contrast with a figure which usually fell between 40 and 45 cm. A selection of age distributions is shown in Table VI. In the case of an unexploited or under-exploited fishery it can be assumed that most of the spawning migration takes place after the modal age has been attained. From the table it can be seen that the Erne eels matured early, at about nine years, and no individuals of 15 or over were found. At the other extreme, in the River Blackwater (the poorest water sampled from a feeding point of view), no fewer than 20 percent of the sample were aged 17 or more and migration did not begin until 14. In the Corrib system migration began between 10 and 12 years, though slow-maturing individuals were plentiful. The situation in the South Basin of Lough Corrib was complicated because, besides the downstream migration at maturity there was a loss by migration upstream toward the higher waters of the system.

In Lough Gill, where large eels were very scarce(Table V), the presence of substantial numbers of eels of 15 years and older suggested that the scarcity was due to intensive fishing rather than to early maturing at a small size. In the case of the South Sloblands the effect of the 1971 fishing on the stocks was clearly shown by the sharp decrease in specimens of more than eight years old.

3.7 Local movements

The scarcity of large eels in the catches made in early summer has already been mentioned, as has the selection of large individuals by the first night's fishing. Two experiments have been made on the effects of using a fyke net over a protracted period. In the Blackwater estuary catches from a net set in a single position had an irregular pattern. Catches per net day were relatively high from 25 June to 4 July, the highest being 11.0 on 2 July. They then remained low but rose to 3.6 on 16 July. Daily fishing was abondoned on 19 July. When it was resumed on 8 August the catch totalled a maximum of 33 and remained high for the next seven days. Following another break there were two very low catches (0.1) on 23 and 24 August, rising to 5.5 on 8 September. The irregular pattern showed clearly that the eels were engaged in extensive wanderings.

In contrast, a similar experiment in Lough Key yielded a maximum catch of 2.25 on the first night, 4 July. Netting continued daily for one month and was repeated for three days after an interval of 22 days (Table VII). No eels were caught on most of the nights after the first and the next highest catch was only 1.3. The mean length of eels taken in the first four days of the experiment (52.92) was significantly higher than the mean of eels taken in the later period (46.6). The decline in numbers of larger eels was not so dramatic as that seen in the Blackwater River but it was interesting to note that the eels which moved into the area in the period after the first week of the experiment were small ones. At the end of the following April a catch of the same order as that for the first night was made. It appears, therefore, that figures for catch-per-unit of effort should preferably be based on the results of single night of fishing in each locality.

3.8 Food

It is not proposed to go into any detail of the feeding habits of Irish eels. It has, however, been suggested frequently that food is digested by eels so rapidly that the stomach contents of net-caught eels are of limited value in studying their food. No evidence of this has been found in these experiments and the stomach contents have been found to be indistinguishable in their state of digestion from those of eels caught by electrical fishing.

About one third of the eels examined after one night's fishing had empty stomachs. The hind guts of half of these were also empty, showing that fasting had taken place for well over twenty-four hours. The examination of the stomach contents of the remaining two thirds usually showed that a sample of thirty specimens or more would give a clear impression of the variety of food organisms regularly eaten. Distinctions between fish-eating and invertebrate-eating eel populations were observed. Information on the size at which fish-eating began was obtained. This varied greatly between populations. There have been indications that the age and hence size at maturity are influenced by the diet. When this has been more fully studied it may be used in assessing the potential of virgin or under-stocked waters as eel habitats.

3.9 Other fish species

The summer fyke net is a remarkably specific fishing gear. In a catch of 3 237 eels in 12 lakes only 180 other fish were captured, namely:

139 perch     Perca fluviatilis
34 pike     Esox lucius
7 rudd     Scardinius erythrophthalmus
3 gudgeon     Gobio gobio

In estuaries many flounder Platichthys flesus and occasional small trout Salmo trutta and salmon Salmo salar, mullet Crenimugil labrosus and bass Dicentrarchus labrus were taken. In the South Sloblands, rudd were caught in plenty and apparently belonged to a very dense population. In the freshwater portion of the River Blackwater, small trout, roach Rutilus rutilus and dace Leuciscus leuciscus were frequent, young salmon were scarce. In an experiment in the River Corrib a train of nets was set in the path of migrating salmon smolts: 3 600 smolts passed within 6 m of the net and two were caught. The nets are therefore virtually useless for the study of fish other than eels but conversely can be used without risk of damage to other fishing interests.


Two habits of the eel, nocturnal activity and propensity for burrowing, make it a difficult species to capture by conventional methods. The extremely slow and irregular growth rate which it shows, at least in temperate waters, tends to inhibit the application of the mathematical models which are normally used in studying exploited fish populations, these being based mainly on homogeneous stocks of species which show rapid and regular growth. Information on eel populations resident in lakes and large rivers is therefore scanty.

A series of surveys beginning in 1965, have been in progress in Ireland using the small fyke net. River estuaries were the first waters to be investigated but the emphasis was soon turned to lakes and later to the larger rivers. It was found that the fyke net operated efficiently in all these habitats and that interesting and useful results could be obtained. Because of the shortage of manpower these experiments had always to be made on a small scale. The important point is that it has been shown that, with limited resources and by collecting very small samples, information of great value to fishermen and fishery owners, to say nothing of its scientific interest, can be obtained.

A catching unit of eight to ten nets can be carried overland by a single person and operated from a boat of as little as 4 m in length. Using one or more such units a sample of about one hundred eels may be collected in a reasonable time. Experience has shown that a sample of this size is adequate to provide the required basic information on food and on age and length distribution, leading to conclusions on growth rate and age at maturity. With experience, based on a number of surveys under a wide variety of conditions, conclusions on the density of population can be drawn. Presumably absolute figures for population density could be calculated by tag and recapture methods if large samples were readily obtained. Such large samples have not yet been sought in these experiments.

A number of precautions must be taken in using the technique. As a result either of hibernation or of seasonal migration, fewer and smaller eels are caught early in the summer than later. In Ireland, by confining the survey to the months of July and August these anomalies were eliminated. The population density, and sometimes the size distribution, varied according to the depth of the lake. In most cases in Ireland the highest concentrations were met in shallow water and primary sampling was confined to depths between 1 and 3 m. Local variations in population density and in size distribution were frequently observed and it was therefore found best to use many short trains of nets distributed over a wide area. A strong tendency was observed in running water for most of the large eels to be captured on the first night of fishing. For effective comparisons of population density the nets should be emptied daily and moved to other positions.

Table I
Effect of different lengths of leaders (Chi square test to compare numbers of empty traps with traps containing one or more eels)
SiteDateLeader length (m)Empty trapsTraps with eelsNumber of trapsEels per trapSEChi square (1 df)
Lough Corrib7–16.5.689.0     7    5   120.420.0316.05
6.5     87    5   920.050.02
Lough Corribwhole
season 1968
9.0     55   21   760.400.101.08
6.5  2791434220.820.04
6.5     13   15   280.890.186.12
4.7     22    6   280.250.10
Lough Gash19734.7     4  10   142.070.511.32
3.3     7    7   141.070.39
Erne Lakes19724.7     53   27   800.540.102.08
3.3     44   36   800.660.10
Blackwater River19734.7     23   35   581.620.231.32
3.3     17   41   582.240.26

Table II
Effect of direction of current (Chi square as in Table I)
SiteOpening of net, facing:Empty trapsTraps with eelsNumber of trapsEels per trapSEChi square
Blackwaterupstream   854621.940.180.9

Table III
Catch-per-unit of effort, Lough Corrib, North Basin 1967–68 Leader length 6.5 m
(net days)
Total catchCatch-per- unit
Dooros Bay3 May–18 June    1–4   92   320.34
24 June–30 August    1–4   23   361.56
3 May–18 June   4–10   31   150.48
24 June–30 August   4–10   50   591.2  
UrkaunmoreAugust 196810–16   12   282.3  
Inishshanboe EastJuly 1967     5–7   721221.7  
Inishshanboe SouthJuly 1967     4–6   803284.1  
Totals17 June–25 August 1967  2213921.77
3 May–18 June 1968 152   750.49
24 June–30 August 1968 1923531.83

Table IV
Catch-per-unit of effort, July and August. Leader length 4.7 or 3.3 m
(net days)
Catch-per- unit
Lough Corrib    
South Basin1969   78   349   4.4
Lough Mask1969106   127   1.2
Lough Carra1969142    71   0.5
Lough Gill1972104   122   1.2
Erne Lakes 1 192   255   1.3
Erne Lakes 2 128     57   0.4
Lough Key1972206   287   1.4
River Blackwater    
Estuary19654421 900   4.3
Careysville1972   18    141   7.8
Careysville1973    8     78   9.8
Banteer1973   11     50   4.6
South Sloblands1970   26   40815.6
 1972   50     15   0.3
 1973   96    455   4.7

Table V
Length distributions, in centimetres to nearest decimal downward (percentage of n)
Lough Corrib         
North Basin     5.68   54542   7   1         7339.60.83
  7–8.68   4394011   6       61543.10.42
Urkaunmore     8.68   020163232         3554.92.76
South Basin     7.69   2433811   6       34941.30.87
Lough Mask     8.69   1205615   8       12746.50.79
Lough Carra     8.69   0174730   6        7147.60.98
South Sloblands     8.70   1315410   4      40844.10.42
      7.73   14647   5   1       45740.50.20
River Blackwater         
      8.65   4383614   8    1 14342.50.29
Estuary     5.66245115   6   4       31135.90.55
River Fermoy15.8.72   028212625        9949.91.3
 16.8.72   17211   511        7539.71.3
    9.8.73   040161727      11649.71.4
 11.8.73   071   61112        3541.22.0
River Banteer10.8.73   016243228        5052.41.7
Lough Gill    06128   6   5       12440.60.69
Erne Lakes    018382420       27249.50.64

Table VI
Age distribution in years of freshwater life (percentage of n)
Lough Corrib        
South Basin  213322516   8   4 331
North Basin  1  82125201015149
Lough Mask  0  5312918   710116
Lough Carra  0  7273225   7   2  68
South Sloblands        
197010302220   9   5   4   91
1973 33a4115   7   2   2   0   66
River Blackwater0  12016251820   89
River Fermoy        
Lough Gill  0  6383513   6   2   95
Erne Lakes  6363718   3   0   0223

a includes two four-year specimens

Table VII
Catch-per-unit of effort and mean length of eels in single position, Lough Key, daily fishing to 5 August
DateCatchMean length
    4 July2.350.4
6 "0.557.7
7 "0.356.5
8 "0.558.3
11 "1.347.9
14 "0.149.3
15 "0.149.3
20 "0.653.8
21 "0.548.3
       4 August0.141.5
no fishing for 22 days  
27 August0.345.0

Previous Page Top of Page Next Page