T. Backiel
Inland Fisheries Institute
05-500 Piaseczno, Poland
The book is addressed mainly to Fishery Biologists but it is hoped that Fishing Gear Technologists also can acquire some basic knowledge of sampling problems and procedures which, in turn, can result in improved design of gears.
To successfully utilize these guidelines the reader should be acquainted with the general biology of fishes, with limnology and elementary statistics. Preferably he should have some knowledge of fishing gear.
The scope of these guidelines is limited to various techniques for capturing and observing fish at all stages of life. It is confined to inland waters, but a wide range of habitats in which fish can be sampled is recognized. Some of these techniques do not differ from those used in commercial or artisanal fisheries, but for sampling they should be used with understanding of the intricate relationships between fish and the gears in operation. The statistical aspects of sampling form an important part of this understanding. The other part is that related to construction and mode of operation of a gear. The reader should be able to choose a gear, or a set of devices, most suitable for his sampling programme out of those described here, although the guidelines do not offer detailed information on the construction of the gears.
In most cases, a gear technologist, or an expert on electric fishing or acoustic methods, should be at least consulted, and sometimes involved in planning and performing research with the selected gear.
Neither the text nor the references cover all sampling techniques and problems related to the design of sampling procedures. Lagler (1971) distinguished several categories of methods of fish capture in two groups: (i) Handling, which includes removal of water, use of chemicals, spearing, netting and electric fishing; and (ii) Sensing, i.e., watching and using photography, television, acoustic methods, etc., all techniques by means of which fish can be located and even counted. The ingenuity of researchers has been far more diverse than the scope of these guidelines.
These guidelines cannot replace actual research experience; nor can reading the chapters below and the literature referred to substitute for operational experience in fish sampling (Lagler, 1971).
Professional fishermen have usually the greatest experience in catching fish and, as mentioned above, the techniques most often employed for fish sampling are the same as those used in commercial fisheries, but because the purpose is to catch fish according to a programme designed with a certain objective in mind, the use of the gear may differ. Commercial fishermen want to catch as much fish as possible and in the most profitable way, whilst a sampling programme has rather different objectives. Thus, the use of the commercial (as well as sport) fishing gears lies within the scope of this book.
The use of commercial (or sport) catches, whether recorded or not, is a somewhat different matter.
This distinction between sampling fish with commercial (sport) gear and using catches obtained and recorded by fishermen is important but an existing fishery is a very valuable, first-hand source of data about the fish stock. If records of catches are available for an extended period, valuable information can be drawn therefrom. Wherever these records are supplemented with data on fishing gear (their kinds, numbers and time of use) inferences on the fishing effort can be made, thus enabling assessment of catchable fish stocks. Methodology employed in these kind of studies lies beyond the scope of this book and can be found elsewhere (e.g., Ricker, 1975).
Statistics of this kind for inland waters are seldom available, however. Notable exceptions exist such as the data on the Polish lake fisheries analysed by Leopold et al. (1975; 1975a,b,c,d). These statistics which record the catch of several fish species by various kinds of gears in identified lakes do not need to be “sampled” and they do not provide information on attributes of fish (e.g., average size of specimens) nor provide samples to be treated in order to assess some attributes (e.g., age distribution). But sampling catch, i.e., taking a catch of fish landed by fishermen can provide material for such assessments. This kind of sampling, which is usually much less expensive than employing one's own gear, should be taken advantage of wherever a commercial fishery exist. But the researcher should be aware that in this case he samples catch but not a fish population in the biological sense. If he knows the techniques used by fishermen and if he is able to assess the limitations of the techniques, his samples can be as good as those collected by means of his own gear and far more abundant for the same effort spent on sampling. The guidelines below also provide general information on how to assess such fishing techniques from the point of view of sampling.
Whatever the objective of sampling fish and whatever the sampling scheme (programme) one should try to evaluate the technique employed. This is partly a problem of bias and variance of estimates as dealt with in Chapters 2 and 12. On the other hand it is a problem of selectivity and efficiency of the technique itself.
Selectivity is a quantitative expression of selection traditionally understood as selection by size of fish (Hamley, 1975) although the reader will find in the chapters to follow the term of “species selectivity” of a gear. The latter is simply understood as relative proportions of numbers or masses of different species captured by a specified gear with a notion that these proportions differ from those in the water body. Obviously, a part of species selectivity can be attributed to the behaviour and distribution of fishes in space and the distribution of samples. Sampling with fyke-nets set in the littoral zone of a lake gives a different species composition than sampling with the same gear set in, say, an outlet from the lake. But there are some really specific reactions apart from the distribution, e.g., those observed in the electric field (see Chapter 7).
Selectivity by size is much better understood (see Chapters 4 and 6) especially with respect to gillnets (Hamley, 1975). As all members of a population are not equally vulnerable to any given method of sampling, selectivity is a quantification of the selection which results from any process that causes the probability of capture to vary with the size of fish. Any sampling programme should include assessment of gear selectivity, but this has very seldom been practised.
The guidelines below do not advise how to perform such an assessment and an appeal for more selectivity studies as suggested further on, which applies not only to gillnets, does not suffice. The reader should consult appropriate literature (Hamley, 1975; Ferguson and Regier, 1963). The best way to do it seems fishing (observing) a known population, as for example marked (but preferably not tagged) fish. Comparison among fishing gears which differ in, say, mesh size, fishing the same population also gives a key to the assessment of size selectivity.
A note seems appropriate here. In gillnets a typical selection curve, i.e., a plot of selectivity factors versus length of fish, is bell-shaped, falling to zero on both ends (see Fig. 4.3). Marine experience with trawls the cod-end of which was covered by a small-meshed bag showed that selection curves are S-shaped. It is, therefore, assumed that above a certain size all fish are captured with the same probability. However, this does not seem true as larger and faster swimming fish escape in front of a towed net at a greater rate than the small ones. This may apply especially to slow moving gear such as seines (see Chapter 6). Unfortunately, we do not know of any selectivity studies on inland water seines based on rigorous treatment of the subject. The same applies to entrapping gear (see Chapter 5). The selective action of electric fishing (Chapter 7) and of chemical methods (see Chapter 8) has been recognized, but size selectivity of these techniques can probably very with technical alterations (e.g., kinds of electric current, e.g., recovery of fish after electrocution or poisoning) and operational skill and conditions.
Selectivity, as defined above, links itself with another often used term, efficiency. The link can be simply explained in that the least selectively captured fish are taken most effectively by a gear. However, efficiency (or effectiveness) is commonly expressed in terms of numbers or mass of fish captured by a fishing gear in a unit of time. A more rigorous definition is that it is percentage removal of fish from a stock, and still more rigorous are the terms “rate of exploitation and catchability” (Ricker, 1975). In these guidelines efficiency is meant either as catch per (some) unit of effort or percentage removal.
While considering the assessment of fish populations, and in particular their abundance, by means of several techniques one wonders how to compare samples obtained with various gears. This problem was tackled by an EIFAC Intercalibration Exercise, the results of which have not been used in the chapters which follow because the report arrived too late. However, even if it came sufficiently early, it would have been difficult to incorporate the estimates and conclusions in these guidelines. On three lakes in Finland, seven kinds of fishing gear were used and 13 species were recorded in all catches. However, only two of them (Rutilus rutilus and Perca fluviatilis) appeared in sufficient quantities in the catches to form some basis for comparison of gear efficiency. The latter was calculated as percentage removal either of Rutilus or of Perca stocks, which were estimated with fair accuracy. For anybody acquainted with commercial catches on lakes the estimates of efficiencies are striking and confusing. For example:
Taking the efficiency of three fyke-nets used over six days in one lake as a base level (1), then
The efficiency of six bottom seinings was 1.5
The efficiency of four trammel nets used six times was 35.5
Now, Leopold et al. (1975) report average daily catches in Polish lakes out of which the proportions are as follows:
| • Fyke-nets (two types) | -1 |
| • Summer seines (of not much larger dimensions than the above) | ca.123 |
| • Trammel nets | ca.6.5 |
Trawls, reported quite efficient in Chapter 6 here, appeared to be rather inefficient during the experiment in Finland. The figures well illustrate the difficulties in the evaluation of fishing, and hence, sampling efficiencies.
Another important result from the intercalibration experiments is the estimates of variability of catches by the gears used there (see also discussion on sources of variability in Chapter 12 here). Firstly, the data confirmed previous findings that the means and variances of the catches were correlated and henceforth, for further statistical treatment the catches were transformed into logarithms. Secondly, that variances were very large, especially with catches by active gear hence, reasonable means can be obtained from a large number of hauls, say 500, which is clearly not possible in small lakes.
From this brief discussion one can say that the sampling techniques described in the following chapters, some of which may sound optimistic, are not easy to evaluate. But the difficulties are mainly associated with sampling schemes aimed at abundance estimates and not so much with sampling for attributes.
Some of the evaluation problems and difficulties mentioned above should not discourage researchers, and I am sure they will not do so as most research workers show “insatiable apetites for data supplied by a large variety of ingenious sampling devices” (quoted from Chapter 10 by Ruggles).
Bagenal, T.B., 1978 Report on the EIFAC/EVO fishing gear intercalibration experiments. Draft produced at Windermere Laboratory of the Freshwater Biological Association, Ambleside, Cumbria, England
Ferguson, R.C. 1963 and H. Regier, Selectivity of four trawl cod ends towards small. Trans. Am.Fish.Soc., 92:125–31
Hamley, J.M., 1975 Review of gillnet selectivity. J.Fish.Res.Board Can., 32(11):1943–69
Lagler, K.F., 1971 Capture, sampling and examination of fishes. IBP Handb., (3):7–45
Leopold, M. 1975 and B. Dabrowski, General premises and selected elements of a method of estimating fish stocks and populations in Polish lakes. EIFAC Tech.Pap., (23) Suppl.1, vol.2:722-7
Leopold, M. et al., 1975 Effectiveness of seine catches for estimation of fish populations in Polish lakes. EIFAC Tech.Pap., (23) Suppl. 1, vol.1:49–57
Leopold, 1975a, Effectiveness of gillnet catches as a tool for the estimation of fish populations in Polish lakes. EIFAC Tech.Pap., (23) Suppl.1, vol.1:90–5
Leopold, 1975b, Effectiveness of trammel net catches as a tool for the estimation of fish populations in Polish lakes. EIFAC Tech.Pap., (23) Suppl.1, vol.1:117–21
Leopold, 1975c, Effectiveness of catches from various types of trap nets for the estimation of fish populations in Polish lakes. EIFAC Tech.Pap., (23) Suppl.1, vol.2:519– 29
Ricker, W.E., 1975 Computation and interpretation of biological statistics of fish populations. Bull.Fish.Res.Board Can., (23) Suppl.1, vol.2:519–29
