Local variability, human factor or vague procedure? Searching for the reasons of excessive mortality in free living aquatic insects, resulting from biological monitoring

During the analysis of the biomonitoring data of Polish watercourses, it was found that the vast majority of aquatic insects (83–95% of more than 8.8 million individuals collected between 2012 and 2019) die unnecessarily during sampling, as a result of the sub-sampling procedure. Great, in some cases 30-fold differences in the mortality of insects, including those belonging to protected taxa, were observed between the samples analyzed by various regional laboratories. The most important reason for these differences appears to be factors that can be collectively referred to as the ‘human factor’, related to removing insects from the sphere of ethical protection, makes their killing an activity ethically neutral. This kind of attitude, which does not attach any importance to the life and suffering of insects interacts with the lack of precise requirements for the sampling equipment used. Those factors seem to be the most important in explaining the causes of excessive mortality of aquatic insects during sampling and the huge differences between laboratories. Principal component analysis and the analysis of variability measures showed that ecological differences between sampling sites can only be considered as a secondary determinant. Improperly planned and conducted monitoring procedures, through excessive mortality of free-living aquatic insects, have or may have a negative impact on the environment and biodiversity. Therefore, it seems very important to promote alternative solutions that reduce such mortality in biological monitoring, such as: non-lethal methods of animal sampling, the use of pre-sampling areas to estimate the density of animals before sampling and estimation of the maximum abundance of a sample necessary for a reliable estimate of taxonomic richness using rarefaction.


Introduction
The degradation of freshwater environments and the resulting reduced biodiversity is one of the most severe type of nature degradation on a global scale (Cazzolla Gatti 2016). According to the requirements of the Water Frame Directive of European Parliament (Directive 2000), quality assessment of freshwater environments in the EU Member States is carried out on the basis of the presence, abundance, diversity and taxonomic composition of aquatic organisms with an established indicative value (biological elements). Aquatic invertebrates ("macroinvertebrates") are the most frequently used group of animals for these purposes. Insects are, in turn, most important and most numerous group of invertebrates used in biological monitoring of watercourses, containing over 70% of macroinvertebrate species in Poland and in Europe (Bogdanowicz et al. 2008). Insects constitute 64% of all collected animals and 62% of all analyzed families in databases used for the purposes of biological monitoring in Polish watercourses (Tończyk and Siciński 2013). Larval forms of certain insect orders are classified as the group with the highest indicative value (so called EPT: Ephemeroptera, Trichoptera and Plecoptera). Many freshwater species of insects in Poland and Europe are endangered (Czachorowski and Buczyński 2000) and particularly sensitive to habitat change as flow alterations, habitat fragmentation and pollution being their main threats (Buczyńska and Buczyński 2019;Sánchez-Bayo and Wyckhuys 2019). Intensively studied the recent worldwide decline in the number and biomass of insects (Hallmann et al. 2017;Leather 2017), widely discussed by specialists (e.g. Montgomery et al. 2020) is undoubtedly linked also to the degradation of freshwater environments. The problem of the observed changes in the number of insects in freshwater environments on a large geographical scale and its reasons are controversial, resulting from the use of different methods of data analysis (Van Klink et al. 2020;Desquilbet et al. 2020). Regardless of this, however, the decline in the population size that has intensified in recent years and the threat of extinction of many ecologically specialized, rare species of aquatic insects seems to be common.
Perhaps large-scale monitoring in which the number of collected specimens reaches hundreds of thousands and millions (e.g. Escribano et al. 2018;Karlsson et al. 2020) may be taken as a potential factor playing a role in reducing the number of free-living insects. The impact on common and abundant species seems to be minimal, as only a negligible proportion of natural populations and habitat area is subject to routine monitoring (Gezon et al. 2015). However, the impact on the populations of rare, local and sparse species seems to be worth considering. The discussions of specialists on the development of standards and good practices in insect sampling for the purposes of monitoring in the face of their global decline began only recently (Samways 2019;Montgomery et al. 2021).
It is very difficult to obtain data on the number of animals killed due to biological assessment because typically only the number of animals included in the analysis is published and the exceptions are very rare (e.g. Escribano et al. 2018). Many biomonitoring procedures used in EU countries recommend the use of sub-samples (Friberg et al. 2009), which results in a significantly higher mortality than this necessary for the analysis. In order to reduce the risk of gaining too little data, researchers often kill far more animals in each sampling effort than necessary to conduct a proper assessment of the ecological status of the habitat. In some situations this type of procedure may take the form of mass killing. A critical approach to this subject is practically not mentioned in the scientific literature.
The arguments against the mass killing of free-living insects in freshwater ecosystems in times of their global decline can be grouped into three categories-the assumptions and ensuing consequences are briefly presented below.

Declining number of insects have negative effects on ecosystems
Populations of many insect species are essential for the functioning of freshwater ecosystems, especially watercourses (Suter and Cormier 2015). These organisms filter a huge amount of edible particles suspended in water, control prey populations as predators, constitute the food base of fish species and consume a periphytic algae and dead organic matter, which accelerates the biomass turnover. The activity of certain groups of aquatic insects reduces the effects of eutrophication, and intensify the self-purification processes in watercourses (Allan and Castillo 2007). Huge swarms of winged mayflies, true-flies and caddisflies after their emergence constitute an irreplaceable food base for many terrestrial vertebrates (Suter and Cormier 2015). As significant reductions in their abundance may cause disturbances in the functioning of the ecosystem, though killing them should be considered a kind of environmental degradation. However, the mortality related to biological monitoring is not able to reach an intensity to cause a decline in common insect species. This type of impact can only be expected in specific environments or directed into isolated populations living in low densities.

Declining populations of wild insects could mean reduction in local biodiversity
This is especially important for rare and protected insect species. Out of approximately 2500 species of aquatic insects found in Poland, 254 species are covered by various forms of legal protection. Many species occur locally, in low densities and are highly ecologically specialized. Particularly, some groups of aquatic insects inhabiting fast flowing waters and sandy bottom of lowland rivers are particularly vulnerable to extinction due to water pollution and hydro-morphological transformation of the environment (Czachorowski and Buczyński 2000;Buczyńska and Buczyński 2019).

Mass killing of insects means suffering and death to many sentient individuals and is therefore unethical
Contrary to widespread belief and contrary to the traditional approach of many scientists, insects-animals with a complicated nervous system can feel pain and display relatively complex modes of perception and cognition (Elwood 2011). Sampling activity involving death of all inhabitants on a fragment of the stream-or river bottom with chemical preservatives, especially when only a small part of them is used for analysis and the rest is killed inadvertently raises serious ethical doubts. Excluding insects outside the scope of moral protection, has been criticized by ethicists and philosophers as unreasonable and distant from modern scientific view (Broom 2013; Mathers 2019) but seems very difficult to be changed by social consensus (Kellert 1993;Simaika and Samways 2018). Several ethical concepts, widely discussed in scientific literature and based on the results of experimental studies, postulate that arthropods, including insects, should also be protected against unnecessary suffering and death (e.g. Mikhalevich and Powell 2020; Baracchi and Baciadonna 2020).
The main reason to take up the subject of this article was the observation of huge differences in the numbers of animals killed during sampling for the purposes of biological monitoring in Poland. There were clear differences in the abundance in samples collected using identical methods and according to the same detailed protocols, both between different types of watercourses and between individual regional laboratories. Most importantly, however, clear and repeatable differences in the number of animals killed during the assessment were found in the database between samples collected on the same type of environment by staff from different laboratories. The preliminary analysis of the database presented in this article showed that the estimated mortality of insects in some samples, exceeds the number of individuals necessary for the proper conduct of the analysis 50 and even over 200 times.
The aim of the analysis was to explain: (1) what is the real extent of the differences in mortality between samples, (2) what these differences result from, (3) what is the level of overkilling in insect mortality as a result of biomonitoring activities. The answer to the first of these questions might emphasize: (i) local differences in the density of fauna between groups of watercourses, resulting from natural causes, e.g. ecological or hydrological parameters; (ii) the specificity of individual laboratories performing biomonitoring assessment, possibly related to "human factor"-the way teams and individual employees work and the interpretation of procedures of sampling and segregating material. The importance of ethical sensitivity in the approach to killing insects was also taken into account and discussed later in the article.

Methods
Data on the abundance of aquatic insects found in 2880 samples of invertebrates collected in watercourses throughout Poland and identified to the level of family were analyzed. The results include the numbers of insects identified to the level of family in sub-samples randomly selected from multi-habitat samples. Determining the ecological state of the environment at a given site on the basis of data on invertebrates requires the implementation of a detailed methodology RIVECO (presented in detail by Bis et al. 2013), which includes the following activities: (i) Selection of the site, as assigned to the appropriate abiotic type; (ii) Quantitative sampling of invertebrates at the site, preparing 20 sub-sampling from the stream or river bottom covering a total area of 1.25 m 2 , using specialized equipment dedicated for this purpose, and taking into account the proportions of all types of substrates present there; (iii) Preserving the samples using a suitable substance which kills all organisms; (iv) Random selection of sub-samples from the preserved organisms covering 1/30 of the total sample. The procedure requires at least five sub-samples to be included in the analysis; (v) Identification of animals selected in sub-samples to the level of family; (vi) Calculation of the value of the multi-metric index MMI on the basis of taxonomic composition and richness of the invertebrate fauna using the dedicated computer application; (vii) Determination of the class of the ecological status on the basis of the MMI final score depending on the habitat type.
This activity in Poland is carried out by network of laboratories of environmental monitoring under the Regional Inspectorates of Environmental Protection. The analyzed database includes the results carried out in the years 2012-2019 by 49 laboratories, each of them represented by 16-209 samples. To ensure impartiality, the data on laboratories have been coded for analysis (double-blinded) and in this paper only their coded numbers are given to prevent their identification. Additionally, for testing purposes, samples of aquatic insects collected at 27 sampling sites by the Author of this work and his colleagues were included into database-these were coded as laboratory 30.
The database provided also includes the following data: According to the recommended procedure (Bis et al. 2013), in order to correctly calculate the value of MMI, data on at least 350 representatives of macroinvertebrates should be entered into the calculations. Among them, on average, there were 217 (SE = 17.63) individual insects, therefore this number was considered the minimum number of insects in the sample, sufficient to correctly assess the ecological status on the basis of a single sample. For each sample, the following was determined: -The number of insects in sub-samples of each sample; -The number of families containing protected species (FCPS) in sub-samples of each sample; -The number of specimens belonging to FCPS in subsamples of each sample.
Based on the number of random sub-samples selected for each sample, it was determined: -The estimated number of insects killed during each sampling effort (ENIK); -ENIK belonging to FCPS; -ENIK killed, but not analyzed (IKNA); -The estimated insect family richness and estimated FCPS richness in each sample obtained by the rarefaction procedure (calculated with EstimateS 9 software (Colwell 2013).
The analysis covered: -Differences between samples collected by individual laboratories; -Differences between samples collected at sites belonging to different abiotic types; -Differences between samples collected by different laboratories at sites belonging to the same abiotic types; -Differences between samples collected in different years and in different months; -Differences between samples collected at sites with different geographic location; -Differences between samples collected at sites classified to different ecological status.
Comparison of standard deviation, coefficient of variation (CV) and standard error was used to determine relative importance of local habitat variability and the specificity of individual laboratories in explaining differences in insect mortality. All three parameters have different application requirements and the comparison of data with them gives slightly different information. For comparison between data sets with widely different means, the CV should be used instead of the standard deviation. CV should not be used to assess the certainty of measurement when the number of replicates significantly varies across compared samples. The standard error is an estimate of how far the sample mean is likely to be from the population mean, whereas the standard deviation of the sample is the degree to which individuals within the sample differ from the sample mean.
The Principal Component Analysis, calculated with PAST 3 software (Hammer et al. 2001) was performed to determine the significance of the variables related to the number of ENIK in explaining the overall variance. For this purpose, the values of each variable were log-normalized. In order to remove the significant relationship between geographic location and affiliation of a data point to a specific laboratory, the values of longitude and latitude were standardized within each laboratory.
Quotient Ql has been found to be a reliable indicator describing the standardized mean value of ENIK in samples processed by individual laboratory in relation to other laboratories. It can be calculated as: where mStl is mean ENIK at sites belonging to type t and analyzed in laboratory l, msSl is mean standardized value of mStl calculated for all abiotic types analyzed in laboratory l, mSt is mean ENIK at all sites belonging to type t.
Quotient Qt is, in turn, standardized mean value of ENIK in samples collected in individual abiotic type in relation to other abiotic types. It can be calculated as: where mStl is mean ENIK at sites belonging to type t and analyzed in laboratory l, msSt is mean standardized value of mStl calculated for all laboratories which analysing samples collected at sites of type t, mSl is mean ENIK at all samples analyzing by laboratory l.

Results
In all sub-samples subjected to taxonomic analysis, 1,510,941 individual insects, belonging to 114 families were found. The total number of insects collected and preserved in the samples was estimated, based on the relative number of sub-samples, at 7,703,465 individuals. Among them, 2,817,777 individuals belonged to 73 families, which include species covered by various forms of legal protection (FCPS). Based on the rarefaction procedures, it was estimated that the real family richness in the total samples was significantly higher than that found in the sub-samples (Fig. 1). Most often the difference has been 4-7 families, but in some cases even 20-21 families. This means that taxonomic identification of all killed animals, even six times more numerous than those present in the sub-samples in accordance with the procedure, would enrich the family richness of insects collected by 5-8%, and exceptionally by nearly 40%. The saturation curves of estimated family richness in samples belonging to the same type and analyzed by different laboratories have a very different shape (e.g. 25 and 36 in type 4; 12, 29 in type 6; 8, 31 in type 24, Fig. 1). This indicates a different rate of appearance of rare families Ql = msSl∕mSt, which is surprising in environments which are very similar in terms of ecological conditions-it results in differences in the estimated richness reaching over 40%. Such observations can only be explained by significant differences in sampling methodology between laboratories.
During the laboratory analysis of approximately 240 samples, a special procedure for simplified counting of individuals of families present in samples at very high densities (SCHD), admitted in the RIVECO method was probably used by the staff of laboratories. This procedure allows the registration of only the first 100 identified specimens of families present in subsamples in very high numbers (Chironomidae, Simuliidae, Baetidae). Since neither the facts of using this procedure nor the actual abundance of these families is not archived, it becomes impossible to precisely determine the size of the entire sample. Despite that, it is possible to confirm that the procedure SCHD has been applied and to estimate the actual number of insects in these samples. 1 It follows that the total estimated number of insects killed during sampling, taking into account the effect of procedure SCHD, should be increased by at least 12%, so it was estimated at 8,627,881.
Principal component analysis showed that only two factors: "laboratory" and "abiotic type" are important in explaining the overall variance associated with the total number of insects killed during sampling (Fig. 2). These two factors are orthogonal to each other. "Laboratory" is much more important than "abiotic type" because it determines the first principal component, which explains six times more of the overall variance (79.1%) than the second one (13.3%). Eigenvalues of only the first two principal components exceed 1 (1.95 and 1.35 respectively). The remaining analyzed factors: longitude, latitude, ecological status (quality class), year and month are of very little importance. Very significant differences in ENIK were found between samples belonging to different abiotic types (Table A1) tested with Kruskal-Wallis analysis, p = 5.30E−32). In most of abiotic types on average approx. 1500-3000 individuals were killed during taking each sample, but at sites of types 14 and 15 (upland small and medium-size rivers) even up to 8000 ind. Mean number of sampled and preserved individuals per sample in these types were 13-15 times higher than those at sites of type 3 (streams in Sudety Mountains). Very significant differences were also found between samples collected, prepared and analyzed by different laboratories in: (1) ENIK (Table A2, Kruskal-Wallis analysis, p = 2.32E−35). Three groups of laboratories, which differ significantly from each other in terms of this parameter, were distinguished: (1) the first group (laboratories 12 and 17) with extremely high values (approx. 13,500 ind. per sample), the second one includes 12 labs with values approximately 2-4 times lower, and the third group includes other laboratories with average values approximately 5 to 40 times lower). All three groups of laboratories belong to the homogeneous groups in the post hoc analysis with the Tukey test. Mean values of Ql in these groups of laboratories were 3.84 (group 1), 1.63 (group 2) and 0.74 (group 3), respectively.
The differentiation of Ql values between particular laboratories (Supplementary Materials, Table 2) was much greater than the differentiation of Qt values between particular abiotic types (Supplementary Materials, Table 1). It clearly follows that the differences in the mortality of insects between individual laboratories are clearly greater than the differences between abiotic types.
The highest values of ENIK were associated mainly with only seven laboratories and were observed in many types of watercourses. In total, the samples from 7 laboratories: 17, 9, 12, 3, 24, 31 and 33 include 39 out of 46 samples (85%) with a mortality of over 20,000 ind. (Fig. 3).
The most striking are the differences between samples collected by different laboratories at the sites classified as the same abiotic types. They concern: (1) In watercourses of type 14, the differences in mean ENIK values between particular laboratories exceeded 10 times, in watercourses of type 24 it exceeded 14 times and in the watercourses of type 17 even 32 times. In these cases, values of ENIK exceed 60-85 times the average number of insects necessary for the correct analysis of the sample for monitoring purposes.
(2) ENIK which are not subject to analysis, it means that they are killed inadvertently (IKNA). The differences The smallest circle represents one and the largest circle five samples between laboratories exceed 9 times (within type 14), 11 times (type 24) and 50 times (type 17). (3) ENIK belonging to FCPS (Fig. 4). These numbers in some cases exceed 5000 ind., and in most cases 83-95% of them are killed unnecessarily because they are not included in the analysis. In the watercourses of type 12, the differences in the average number between laboratories exceeded 23 times, in type 6 32 times and in type 24 even 115 times. These differences were greatest within types 6, 14, 16, 17, 24. Five laboratories: 3, 9, 12, 17 and 31 were responsible for the killing of a total of 64% of all analyzed FCPS individuals (estimated at 1,797,811).
The comparison of various measures of variability between groups of samples (coefficient of variance, standard deviation and standard error) showed ( Table 1) that values of ENIK are more diverse when mean results of different laboratories in the same type of watercourse are compared than when mean results obtained by the same laboratory in different types of the watercourses are compared. These differences are statistically significant (Mann-Whitney's test): CV p = 0.0008, SD p = 0.0003, SE p = 0.0035. These results are fully consistent with the results of the PCA analysis (Fig. 2), according to which the differences in the number of insects killed during sampling are determined primarily by the specificity of individual laboratories.

Discussion
The differences between the density of insects in the analyzed database at different sites, belonging to the same abiotic type and collected in the same year by the same laboratory, most often ranged between 3 and 6 times. In some cases, especially in upland watercourses (types 4 and 6), however, they were above 20 and this level of variability  is congruent with other results in similar environments in Europe (e.g. Ortiz and Puig 2007;Kelly-Quinn et al. 2003) and in Poland (e.g. Lewin et al. 2013;Brysiewicz et al. 2020). Such a level of variability should therefore be considered typical, natural and resulting from the ecological and climatic differentiation of sites within the same abiotic type. However, the level of variability in the sampling mortality is clearly lower than that resulting from the differences between laboratories. The mean differences in ENIK values between laboratories concerning the samples collected in watercourses belonging to the same abiotic type, and therefore very similar in ecological terms, clearly exceed the natural variability between habitats and are highly statistically significant.
It is difficult to establish the direct causes of such a great mortality during sampling in some laboratories on the basis of the database. The term 'excessive mortality' used in the title and the abstract of the work should also be considered in more detail. From an ethical point of view, considering the right and wrong dealings with living beings, 'excessive' can mean the death of any number of animals, if only we consider them to be suffering and endowed with inherent moral value. From the point of view of nature conservation, the meaning of this term is more difficult to define. It certainly concerns a situation where the level of animal mortality permanently threatens the stability of the population. In a situation such as that described in this paper, however, it is impossible to even attempt to assess the strength of such an influence. So, it seems justified to use the term 'excessive mortality' to describe the effects of sampling methods in which many thousands of specimens are killed non-selectively, most of which are not used for any purpose. Perhaps the reason partially lies in the sampling methods used, although in my, the most convincing explanation may lie in the broadly understood "human factor".
The RIVECO methodology allows different samplers, but recommends the use of the hand-net, especially in environments with heterogenous bottom substrate (Bis et al. 2013). The analyzed database does not contain information which of the tools was used to collect samples by staff of a laboratory. The data presented by Cunha et al. (2019) suggest that the hand-net is conducive to taking more numerous samples and that other commonly used samplers to collect fewer specimens per area unit. It seems that it is very difficult to retrieve single very small sub-sample in a repetitive manner on stream bottom using the hand-net precisely on the way of about 25 cm. This requires a lot of experience and the procedure and practice are aimed at avoiding the risk of taking too little and not too much sample. The surface of the bottom from which single samples should be taken with the Surber's and Gunther's sampler is, unlike the hand-net, fixed and limited, and therefore more repeatable. Taking a relatively small sample increases the effort and extends the time of the sample processing (you need to select more sub-samples) while taking large sample reduces the effort and time because it requires only a standard 1/6 of the entire sample to be processed. Importantly, taking a small sample also causes the possibility of not reaching the required number of 350 animals, which should lead to the rejection of the sample results. It must be emphasized that the procedure does not use a mechanism limiting the maximum sample size. Hence, the distribution of ENIK values is clearly skewed-in 72% of the samples at least three times more animals than necessary were killed and in 51% more than 10 times (Fig. 5). The aforementioned lack of mechanisms limiting the maximum sample abundance also touches upon ethical issues-reducing the number of animals killed is not a first, or even second priority of any official biological monitoring methods. Benthic invertebrates, with the exception of protected species are still treated as non-sentient, thus devoid of intrinsic value, and therefore do not require any form of ethical protection.
"Human factor" is a set of factors that is difficult to define and to objectively assess. I would like to emphasize that I am far from a critical evaluation of the work and attitude of anyone from the staff of any laboratory. However, removing insects, as well as other aquatic invertebrates from the sphere of ethical protection, makes their killing, also during monitoring, an ethically neutral activity, regardless of the number of animals killed. It should be emphasized that such commonly held beliefs about relations between humans and animals are usually based on a personal worldview shaped by religious principles as argued, e.g., by Barad (1988) andBruers (2014). Probably, for some people participating in this type of activity, obtaining data by counting the bodies of killed animals does not differ, in ethical terms, from the assessment of the concentration of microorganisms or plant biomass in water, or even the concentration of chemical compounds dissolved in water. Assuming the existence of this kind of ethical insensibility, taking an insect sample larger than required may be equivalent to taking an additional water sample and performing an additional analysis of phosphate concentration. It then becomes an act perceived positively. This kind of ethical attitude, which does not attach any importance to the life and suffering of invertebrates, unfortunately also seems to be typical for authors of procedures that do not limit in any way the maximum number of animals killed during the sampling. It seems likely, that the lack of generally accepted ethical standards for killing insects may affect personal attitude related to ethical sensitivity and thus individual behavior (Samways 2010). The priorities underlying modern monitoring procedures, appear now to be obsolete, because they do not take into account ethical aspect of killing sentient individuals. The previously mentioned factors, together with the lack of precise requirements for the sampling equipment used, seem to be the most important in explaining the causes of excessive, unnecessary mortality of aquatic insects during sampling and the huge differences in mortality between laboratories.
Large scientific projects in which insects are monitored for biodiversity assessment also result in significant mortality (Gadagkar et al. 1990-tens of thousands preserved individuals; Duelli et al. 1999-hundreds of thousands;Karlsson et al. 2020-tens of millions). However, they differ from the results discussed in this paper by very important feature. Namely, unnecessary mortality was reduced there to a minimum-all individuals are by definition identified with maximum accuracy by specialists or archived for future identification. The goal is to know biodiversity, so regardless of the number of individuals killed, the death of each of them matters. In the case of the discussed data on the ecological quality assessment, the majority of individuals are intentionally, but unnecessarily killed and their dead bodies are later not used for anything.
This kind of ethical insensibility is also in contradiction with the more and more commonly formulated proposals to grant ethical value not only to individuals but also to ecosystems (Naess 2002) and to biodiversity as such (Wilson 2000;Clausnitzer et al. 2017;Simaika and Samways 2010) as well as the calling for the saving of animal life as an element of "natural offset" (Ives and Bekessy 2015). An increasing amount of evidence from experimental research is revealing that many groups of invertebrates have complex behavior and advanced mental potential: substantial perceptual ability, feelings and emotions, pain perception and long-and short-term memory (Broom 2013), learning abilities and cognitive perception (Perry et al. 2013). One of the manifestations of this may be individual differentiation of behavioural types (Kralj-Fišer and Schuett 2014; Carrere and Mather 2019). Many of these results have been obtained in experiments with aquatic insects and includes: adjusted hunting strategies, stimuli related anti-predatory strategies, complicated mating behavior, kin-recognition and relationships within and between kin lineages (Cárcamo and Spence 1994;Koperski 1998;King 1999). The view that any activity that causes pain or death in sentient animals should be limited and justified only if the benefits to other organisms or the ecosystem as a whole are significant is becoming more and more accepted (e.g., Mikhalevich and Powell 2020;Howard and Symonds 2020).
The development and widespread use of non-lethal methods of analysis eDNA will probably eliminate such ethical dilemmas in the future (Deiner et al. 2017;Fernández et al. 2018). It should be noted, however, that the metabarcoding methods used to assess the ecological condition of freshwater environments (e.g. Elbrecht and Steinke 2019), which have been dynamically developed in recent years, are not associated with a significant reduction in animal mortality during sampling. I recommend some alternative changes to the biological monitoring procedures of freshwater environments presented below which could significantly reduce the mortality of freshwater insects: (i) complete resignation from the use of animals in monitoring and limiting it to the analysis of non-sentient organisms only: microorganisms (bacteria, algae, protists) and aquatic plants. (ii) The development of non-lethal methods of animal sampling, based on high-resolution digital photography (Karasek and Koperski 2015) or the visual survey of adult stages, possible for some large insects (Bried and Hinchliffe 2019; Kietzka et al. 2021). (iii) Use of pre-sampling area taken near the sampling area in order to estimate the density of bottom animals before sampling. (iv) Obligatory application of rarefaction to create empirical saturation curves determined for types of environments, which will allow the estimation of the maximum abundance a sample necessary for a reliable estimate of taxonomic richness.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.