In situ bioindications enable determination of the suitability of an environment for endangered mussel species. We describe two methods based on the juvenile exposure of freshwater pearl mussels in cages to oligotrophic river habitats. Both methods are implemented in variants for open water and hyporheic water environments.
Knowledge of habitat suitability for freshwater mussels is an important step in the conservation of this endangered species group. We describe a protocol for performing in situ juvenile exposure tests within oligotrophic river catchments over one-month and three-month periods. Two methods (in both modifications) are presented to evaluate the juvenile growth and survival rate. The methods and modifications differ in value for the locality bioindication and each has its benefits as well as limitations. The sandy cage method works with a large set of individuals, but only some of the individuals are measured and the results are evaluated in bulk. In the mesh cage method, the individuals are kept and measured separately, but a low individual number is evaluated. The open water exposure modification is relatively easy to apply; it shows the juvenile growth potential of sites and can also be effective for water toxicity testing. The within-bed exposure modification needs a high workload but is closer to the conditions of a natural juvenile environment and it is better for reporting the real suitability of localities. On the other hand, more replications are needed in this modification due to its high-hyporheic environment variability.
The exposure of experimental organisms in situ with the subsequent evaluation of their condition is one possible way to get information about the environmental quality and (especially) the site suitability for a species. Within animals, such a bioindication is applicable primarily for small invertebrates which are able to live in a limited bounded space. Young stages of bivalves (Bivalvia) are one such suitable organism group1.
Bivalves of the family Unionidae are a very important component of aquatic ecosystems2. However, these species are often critically endangered, especially in streams and rivers. Some of them are characterized as 'umbrella species' whose conservation is closely related to the conservation of the whole stream biotope and which require a comprehensive approach3. These animals have a life cycle associated with many environment components, from water chemistry4,5 to changes in the populations of fish which serve as mussel larvae hosts6. Because mussel juveniles often represent a critical phase of the mussel life cycle, the site suitability for their development at this stage is crucial for a successful species population development in a locality.
The freshwater pearl mussel (FWPM, Margaritifera margaritifera; Unionida, Bivalvia) is a critically endangered bivalve occurring in oligotrophic European streams. Their numbers have fallen drastically during the 20th century across the occurrence area. It seems that the current decline in species reproduction in the majority of the central European populations is primarily caused by very low to zero survival of juveniles during the first few years of their life. It is assumed that juvenile FWPMs live for many years in the shallow hyporheic zone7, of which the conditions and their variability are still not well described. Moreover, until their second year of life, the juveniles only have a dimension of up to about 1 mm, so they are very difficult to find in large volumes of sediment under natural conditions8. Therefore, experiments with captive juveniles are necessary for the study of their ecology.
Within the Czech Action Plan for Freshwater Pearl Mussel9, there are thousands of juveniles rising every year from a semi-natural breeding program. Nevertheless, there is a question of which localities and habitats are suitable for successful population support by these juveniles or for eventual species reintroduction. In situ bioindications present a way of finding the answer.
Despite the fact that inconsistent survival rates of juvenile mussels in exposure cages were observed in some earlier works that questioned the suitability of juvenile mussels as bioindicators10, several recent studies have confirmed the applicability of juvenile exposure methods for water quality testing11,12,13. Additionally, it has been demonstrated that several factors need to be considered when interpreting the results of these particular studies, such as the stock origin14 and the persisting effects of larval conditions15.
The question arises of how to install experimental juveniles in tested localities and how to most effectively evaluate their condition. The first rigorous application of in situ exposure methods with juvenile FWPMs was published by Buddensiek16. Juvenile FWPM individuals were kept in sheet cages, exposed in the free-flowing water of streams, and their survival and growth were quantified after several weeks of exposure. The approach was originally developed as a semi-artificial breeding method, but the author also highlighted its applicability for the assessment of habitat requirements and water quality. Although the FWPM juvenile survival is naturally very low on a scale of months/years and only a very small number of animals will survive, the survival rate can be a good marker of the environmental effect on a scale of several weeks16. Over years of research, exposure methods were developed further to hold experimental juvenile mussel in-stream habitats and to evaluate their growth and survival rates; these include sandy boxes17, mussel silos based on an upwelling principle18, and various other exposure cages (summarized by Gum and colleagues)11. Because juveniles occur naturally in shallow hyporheic zone7, the application of experimental devices within the stream bottom is very desirable.
In our article, we describe the use of two exposure devices for FWPMs: i) modified Buddensiek sheet cages ("mesh cages") also enabling bioindication testing in hyporheal conditions; and ii) Hruška sandy boxes ("sandy cages"). The protocol describes the application of both methods in open water and hyporheic conditions (i.e., four variants of exposure are described). The methods were gradually modified and expanded over more than 15 years of application within the Czech Action Plan for Freshwater Pearl Mussel9 and verified by a set of experiments.
1. Mesh Cage
Note: See Figure 1.
2. Sandy Cage
Note: See Figure 2.
The four bioindication methods (open water sandy cages, within-bed sandy cages, open water mesh cages, and within-bed mesh cages) were applied to investigate the environment condition suitability for FWPMs in the upper Vltava River Basin (Bohemian Forest, Czech Republic). This river represents one FWPM residual locality within central Europe19. Here, we present a specially selected set of results illustrating the most important aspects of the four methods. Further details are described in a comprehensive study by Černá et al.13.
The river environment was studied at two levels:
(I) A longitudinal river profile was represented by main stream localities (sites A – E) and tributaries of different pollution stages (sites R and V). The localities were tested both by sandy cages and by mesh cages installed in free-flowing water. In addition, a gravel hyporheic zone was tested by within-bed sandy cages in localities B, C, and D.
(II) A hyporheic environment was tested in the selected locality C. The suitability of different substrates (sand, gravel, stones) was tested by within-bed mesh cages.
The growth rate and survival rate of >1-year-old juveniles (see Supplementary File 1: S.6. Biological material) were tested. The experiment was carried out to its full extent in the summer of 2014 and was repeated to a smaller extent at some localities in the summer of 2015. Within level (I), 2 – 6 sandy cages with a minimum of 100 juveniles and 6 (2014) or 4 (2015) mesh cages with 6 juveniles were applied at each locality tested by the appropriate method. Within level (II), 7 mesh cages with 6 juveniles were installed in every tested environment. The exposure time was one month for the mesh cages and three months for the sandy cages.
The statistical analysis was conducted in R, version 3.1.020. Kruskal-Wallis, Kruskal-Nemenyi, and Wilcoxon-Mann-Whitney tests were used. For data with a normal distribution, linear or quadratic regression was carried out.
The localities can be clearly distinguished based on the growth rate in the open water mesh cages despite the high within-cage variability, even in different growth-favorable periods (Figure 3). In the more growth-favorable exposure in 2015 (growth rate 19.3 – 41.8%), a significant trend was discovered in the longitudinal profile where the growth rate increased downstream (Kruskal-Wallis test, p < 0.001). Importantly, the survival rate was equivalently high in both seasons (from 83%) (Figure 4A).
On the other hand, the open water sandy cages showed a different trend between the main stream localities in 2014: the growth rate increased downstream from locality A (52%) through the middle locality C (153%), and thereafter decreased again until locality E (46%) (a quadratic regression of absolute growth values: r2adj = 0.77, F2,13 = 25.66, d.f. = 16, p < 0.001). This trend was also confirmed in 2015 when the greatest growth rate was recorded at the middle locality C again. Also, the absolute growth rate values did not differ much between 2014 and 2015. On the other hand, the survival rate differed between the years, being much higher in 2015 (from 48% to 72%) than in 2014 (about 25%) (Figure 4B).
An effect of two different exposure methods is also clearly visible in the polluted tributary (locality V). The sandy cages exposed here during the three months showed 0% survival, while an 83% survival rate with some growth was recorded by open water mesh cages exposure here during the 30 days.
Results from the within-bed sandy cages illustrate different conditions in the hyporheic environment in comparison to open water in the relevant localities. The growth rate was always lower in the hyporheal sites than in open water, and the survival rate was much more variable (from almost 50% to 0%, Figure 4B).
A study of hyporheic microhabitats using the within-bed mesh cages showed a significant effect of the substrate composition on juvenile survival. The best conditions were recorded from the oxygen-saturated stony bottom (a survival rate close to 100%) while the worst (a < 40% survival rate) were indicated in poorly oxygenated sand where a very high variability in surviving was also detected. Hyporheic water oxygenation, which was repeatedly measured during the experiment, explains this trend (Figure 5).
Figure 1. Bioindication mesh cage with individual chambers. See Supplementary File 1 for further details. Please click here to view a larger version of this figure.
Figure 2. Bioindication sandy cage. See Supplementary File 1 for further details. Please click here to view a larger version of this figure.
Figure 3. Individual variability in juvenile growth rate recorded by open water mesh cages in localities B and E during two seasons. The means and standard deviation are described for every mesh cage. The values are based on the measurement of 6 juveniles (or 4 – 5 juveniles if the mortality rate > 0%) in every mesh cage.
Figure 4. Example results from a field bioindication experiment with mesh and sandy cages. (A) This panel shows example results from a field bioindication experiment with mesh cages. A total of 6 localities (B, C, D, E, R, and V)within the Vltava River catchment were tested on 2 separate occasions (in 2014 and 2015). The exposure time was 30 days during the summer season. The Localities B – E represent (in order) a longitudinal profile of an approximately 20-km stretch of the river's main stream. Localities R and V represent profiles of 2 tributaries. Capitals mark the same locality both in panel (A) and (B). All localities were tested with open water mesh cages. In addition, locality C was also tested using within-bed mesh cages installed in 3 different types of river bed (Cs = sand, Cg = gravel, Cst = stones) in 2014. The cages were installed in 4 – 7 replications at every site. 6 freshwater pearl mussel juveniles of 1+ year old were used per mesh cage. The average growth rates are marked for the 3 largest individuals (3 MAX) from every tested mesh cage (columns, left axis) and the average survival rate per mesh cage (blue points, right axis). (B) This panel shows example results from a field bioindication experiment with sandy cages.A total of7 localities (A, B, C, D, E, R, and V)within the Vltava River catchment were tested over on 2 separate occasions (in 2014 and 2015). The exposure time was 3 months during the summer season. Sites A – E represent (in order) a longitudinal profile of an approximately 30-km long stretch of the river's main stream. Sites R and V represent profiles of 2 tributaries. Capitals mark the same locality both in this and in the previous panel. All localities were tested with open water sandy cages. In addition, localities B, C, and D were also tested using within-bed sandy cages installed in gavel river bed substrate (Bg, Cg, and Dg) in 2014. The cages were installed in 2 – 4 replications at every site. At least 100 freshwater pearl mussel juveniles were present in every sandy cage. The average growth rate for the 10 largest individuals (10 MAX) from every tested sandy cage (columns, left axis) and the average survival rate per sandy cage (blue points, right axis) are marked. Please click here to view a larger version of this figure.
Figure 5. Oxygen saturation. This panel shows the relationship between minimal values of oxygen saturation over 30 days of mesh cages exposure and the surviving rate per cage in within-bed mesh cages exposed in different bed microhabitats in 2014. Please click here to view a larger version of this figure.
2014 | 2015 | |||
locality | 3-month exposure of sandy cages | 1-month exposure of mesh cages | 3-month exposure of sandy cages | 1-month exposure of mesh cages |
A | 13.9 | – | – | – |
B | 14.4 | 13.4 | 13.9 | 17.5 |
C | 15 | 13.8 | 14.4 | 18.3 |
D | 15 | 13.8 | 14.3 | 18.3 |
E | 15.5 | 14 | – | 18.7 |
R | 13.5 | 12.8 | – | – |
V | 14 | 13.2 | – | – |
Table 1. Average surface water temperature (°C) at the localities during exposure in 2014 and 2015.
Exposure time:
Even one-month exposed mesh cages show a visible growth reflecting differences between localities (Figure 3), so they are very usable for the quick and easy detection of a locality characterization. Nevertheless, the relevance of the results depends on the short-term state of the conditions, which can oscillate. In particular, short rainfall events can play a role. In contrast, unpredictable episodic pollution may not always be recorded. In locality V (Figure 4A), water chemistry analysis detected a short wave of strong ammonium increase13. This was probably responsible for the mortality in the three-month exposed sandy cages but did not affect the 30-day exposed mesh cages.
Temperature fluctuations can also affect short-term exposure results. The one-month average temperature during mesh cage exposure differs between the years (Table 1). The growth rate also varies where higher temperatures were accompanied by higher growth rates (Kruskal-Wallis test p < 0.001). On the other hand, the average water temperature at the same localities during the three-month sandy cage exposure was very similar in both years (Table 1) and the growth rate did not differ significantly (Figure 4B).
Benefits and failings of the described methods:
An open water exposure is relatively easy to perform but is of limited value for habitat bioindication. The method of open water mesh cages is relatively old16 and has been repeatedly used with minor modifications10,11,12,13,21,22,23. Nevertheless, these cages are not limited by oxygen, whose deficiency is probably responsible for many juvenile deaths in hyporheic conditions. Thus, open water mesh cages can show good development even in localities with increased mortality and a declining growth rate in open water sandy cages (locality E) or a 100% death rate in within-bed sandy cages, as at locality D in 2014 (Figure 4B). Apparently, the open water mesh cages show locality growth potential, but this may not be realistic as it is dependent on the real availability of hyporheic microhabitats within a locality. Because the open water mesh cages have the ability of high survival (Figure 4A), even up to a 100% survival rate13, they can serve well for the bioindication of chronic toxicity (or acute toxicity if it is expected at a given time). Also, they can be a useful food source presence testing to some extent.
As a new and uncommon method, the open water sandy cages better simulate hyporheic habitat conditions. Movement of the juveniles between sand grains is possible in this apparatus, which helps to reduce the biofilm growth on the juvenile shell. A hyporheic oxygen deficiency can be caused by the activity of microbes colonizing the sand grains; this effect can also partially occur in cages placed above a river bottom. Nevertheless, due to the necessary periodic cleaning of clogging drifting material from a cage, fine sediments are also removed and thus the conditions are changed in comparison to the natural hyporheic habitat. So, the growth rate can also be considered as locality growth potential in open water sandy cages. However, this is closer to real locality suitability than in open water mesh cages. Therefore, the longitudinal growth rate gradients recorded by sandy cages (Figure 4B) also seem to be more plausible and indicate a more suitable river stretch. Moreover, in sandy cages, the possibility of juveniles and subadults breeding up to sexual maturity is verified9, so sandy cages can serve as a safe breeding and biomonitoring method simultaneously.
Sandy cages and mesh cages placed in the within-bed position are closest to the real conditions in a shallow hyporheal. By allowing a juvenile's movement, sandy cages, in particular, offer them both a vertical and horizontal gradient of several centimeters in scale. This ability to move could be very important for escaping from temporary oxygen-deficient micro-zones. This possibility is absent in the within-bed mesh cages. Therefore, a relatively high number of bioindication units is necessary, because the hyporheic conditions are very variable13,24 (Figure 5) and losses due to an unsuitable location are common.
In summary, the bioindication methods used in this research correspond with presumed juvenile natural conditions in the following order:
1. open water mesh cages,
2. open water sandy cages,
3. within-bed mesh cages,
4. within-bed sandy cages.
The workload per unit increases in the same order. Moreover, the juvenile number required for a statistical testing of the results obtained increases in within-bed exposures too. It seems that the within-bed sandy cages represent a more expensive but accurate bioindication method. This new method needs more testing in the future and comparison with other types of hyporheic studies based on the piezometer measurements25,26. In particular, there is a need to study the degree of similarity using a direct probe measuring physicochemical conditions in the cages and in the surrounding hyporheic environment.
The number of individuals measured in one cage:
Compared to mesh cages, it is not possible to measure the growth increment of specific juveniles in sandy cages as there is no information on which individual from the input set is which one in the output. It is necessary to work with an average value. If counted for all the individuals, this value can be very low due to a number of very slowly growing specimens; however, a couple of individuals can grow very quickly (growth jumpers). Such uneven growth is typical for mussels27. The growth variability among juveniles rises with increasing exposure time and big differences can occur, especially in growth-favorable seasons. Also, a long exposure leads to major mortality rates in the mesh cages (for a review see Lavictoire, Moorkens, Ramsey, Sinclair, and Sweeting28), so we can work with a significantly lower number of individuals at the end of the experiment compared to the input juveniles' set. Measuring only the several best-growing juveniles is a possible method.
The experience of FWPM breeding within the Czech Action Plan for Freshwater Pearl Mussel9,29, as well as the results from experiments on sea bivalves30,31, suggests that growth-deficient juvenile bivalves have a high mortality rate, and there is only a negligible chance of their living to maturity. In contrast, growth jumpers have a higher survival rate and they are crucial for a population recovery. The parameter 10 MAX (the 10 most-quickly growing individuals) takes the growth jumpers into account and can increase the informative value of the experiment, even if high mortality takes place (Figure 4B, season 2014). It should be noted that the growth estimate obtained by this method cannot be a false positive. It may only be slightly underestimated because many of the largest juveniles at the end of the experiment would have grown a little bit more in this case. Also, the workload is less if only 10 individuals are evaluated. Similarly, a measurement of three maximally growing individuals (3 MAX) proved to be appropriate in mesh cages, eliminating the influence of slowly growing, non-perspective individuals, which could bias the real image of site growth potential.
The authors have nothing to disclose.
Michal Bílý and Ondřej P. Simon were supported by grants from the Czech University of Life Science [Internal Grant Agency of Faculty of Environmental Sciences, CULS Prague (42110 1312 3175 (20164236))]. Support for Karel Douda came from the Czech Science Foundation (13-05872S). Data on the bioindication and present occurrence of pearl mussels were collected during the implementation of the Czech Action Plan for Freshwater Pearl Mussels managed by the Nature Conservation Agency of the Czech Republic, which is funded by the government of the Czech Republic and is available at
biological material maintenance and care | |||
Freshwater pearl mussel juveniles | any | NA | from a FWPM breeding programme |
plastic boxes | any | NA | |
thermobox | MERCI | 212,070,600,030 | There are many possibilities. This is one example only. |
field thermobox (ca25 l) | any | NA | cold box (insulated box) commonly used for food transport |
river water | any | NA | |
Petri dishes | any | NA | |
plastic Pasteur pipettes with balloon bulb (droppers) | any | NA | hole diameter 1 mm |
hydrogen peroxide | any | NA | |
plastic container (ca 50 l) for river water | any | NA | |
plastic tea strainer | any | NA | commonly used in kitchen |
mesh cages construction | |||
main plastic bodies | any | NA | |
plactic covers | any | NA | |
special technical sieves 340 µm | Silk &Progress | UHELON 20 T | |
special technical sieves 100 µm | Silk &Progress | UHELON 67 M | |
rubber hose (diameter 5.5 mm) | any | NA | |
steel bolts | any | NA | |
steel nuts | any | NA | |
spanner | any | NA | |
steel spikes | any | NA | |
pliers | any | NA | |
beakers | any | NA | |
plastic dishes (ca. 25x15x3-5cm) | any | NA | |
squirt bottle | any | NA | |
field protocols | any | NA | |
stationery | any | NA | |
plastic container | any | NA | |
string | any | NA | |
hammer | any | NA | |
sandy cages construction and use | |||
sieve 1 mm | any | NA | |
sieve 2 mm | any | NA | |
special technical sieves 340 µm | Silk &Progress | UHELON 20 T | |
plastic boxes with tight-fitting lid | any | NA | |
hot melt adhesive | any | NA | |
plastic box (ca 250 x 150 x 100 cm) | |||
big plastic box (ca 25 l) | any | NA | |
flat stone | any | NA | |
net | any | NA | |
river sand | any | NA | |
round containers | any | NA | |
magnifying glasses | Carson | Carson CP 60 | There ar many possibilities. This is one example only |
cages installation and maintenance | |||
field temperature dataloggers | ONSET | UA-001-64 | http://www.onsetcomp.com/products/data-loggers/ua-001-64 |
spade | any | NA | |
toothbrush | any | NA | |
experiment evaluation | |||
trinocular dissecting zoom stereo microscope | Bresser optic | ICD 10x-160x | There are many possibilities. This is one example only. |
digital camera/ electronic eyepiece | Bresser optic | MikroCamLab 5M | There are many possibilities. This is one example only. |
Calibration gird | Am Scope | SKU: MR100 | There are many possibilities. This is one example only. |
external power source with two movable light guides | Arsenal | K1309010150021 | There are many possibilities. This is one example only. |
Image software | ImageJ software | There are many possibilities. This is one example only. | |
table processor | MS excel | There are many possibilities. This is one example only. |