Reproduction of Pisidium casertanum (Poli, 1791) in Arctic lake

Freshwater invertebrates are able to develop specific ecological adaptations that enable them to successfully inhabit an extreme environment. We investigated the brooding bivalve of Pisidium casertanum in Talatinskoe Lake, Vaigach Island, Arctic Russia. Here, quantitative surveys were conducted, with the collection and dissections of 765 molluscs, on the basis of which analyses on the brood sacs length (marsupia) and the number and size of embryos, were performed. In this study, the number of brooded embryos was positively correlated with the parent's shell length. The number of extramarsupial embryos was much lower than the number of intramarsupial embryos. Our research also showed that the brood sac length and embryos within one individual can vary significantly. Thus, we detected that P. casertanum has a specific brooding mechanism, accompanied by asynchronous development and embryos release by the parent. We suggest that such a mode could result in the coin-flipping effect that, presumably, increases the population breeding success in the harsh environment of the Arctic lake.


Summary
Freshwater invertebrates are able to develop specific ecological adaptations that enable them to successfully inhabit an extreme environment. We investigated the brooding bivalve of Pisidium casertanum in Talatinskoe Lake, Vaigach Island, Arctic Russia. Here, quantitative surveys were conducted, with the collection and dissections of 765 molluscs, on the basis of which analyses on the brood sacs length (marsupia) and the number and size of embryos, were performed. In this study, the number of brooded embryos was positively correlated with the parent's shell length. The number of extramarsupial embryos was much lower than the number of intramarsupial embryos. Our research also showed that the brood sac length and embryos within one individual can vary significantly. Thus, we detected that P. casertanum has a specific brooding mechanism, accompanied by asynchronous development and embryos release by the parent. We suggest that such a mode could result in the coin-flipping effect that, presumably, increases the population breeding success in the harsh environment of the Arctic lake.

Introduction
In the Arctic, where the environmental conditions are extreme (i.e. the lake is frozen to the bottom) for the hydrobionts which inhabit freshwater ecosystems, there is only a short summer season, allowing the growth and reproduction of invertebrates [1,2]. Clearly, the species occurring in the Arctic environment have appropriate ecophysiological and life-history trait adaptations to these harsh conditions [3]. A review of the literature shows that the adaptation ability of invertebrates with respect to habitat in the Arctic has been actively investigated recently.    Talata-Karskaya River [26]. The lake is shallow, with maximum depth not exceeding 1.5 m, and prevalent depth is ca 0.5 m, owing to its thermokarst origin. In winter, the lake freezes to the bottom [32]. The lake has snow-related atmospheric nutrition which it, essentially, enters during intensive snowmelt [26].   a rectangular hand net (0.28 × 0.5 m). At three stations, replicates were performed in triplicate every 10 days (3 August, 13 August, 23 August) in order to study the brooding of P. casertanum [33]. Samples were washed using a hydrobiological sieve (mesh size 0.56 mm) fixed with 96% ethanol in the field and transported to the laboratory for sorting and identification.

Morphology of shells, brood sacs and embryos estimates for Pisidium
We examined a total of 795 of P. casertanum specimens in the laboratory using a stereomicroscope (Leica M165C, Leica Microsystems). All specimens were measured and then dissected in order to record the ratio of gravid animals in each length class and the presence of brood sacs and embryos (figure 2). The estimated stages of sexual maturity were recorded, taking into account the approaches used by other researchers [14,34]. We defined as mature specimens (having embryos), those individuals with shell length of at least 2.2-4.2 mm and the juvenile specimens (lacking the offspring) as non-gravid individuals with shell length of less than 2.2 mm. In total, 139 specimens were gravid, among which 95 individuals had formed two brood sacs. The remaining individuals were not studied because 20 individuals had extramarsupial larvae (i.e. individuals which have broken free from the brood sacs) and 24 individuals had one brood sac at the formation stage. The left and right marsupial sacs were measured with respect to maximum length and then dissected in order to estimate the number of embryos per sac which was used to denote all classes found within an adult. Deviations were calculated for the sac pair of each specimen (n = 95) as D = SL max − SL min ; where SL max and SL min are maximal and minimal sizes, respectively, of each sac pair. Embryos were removed from marsupial sacs and measured using the greatest dimension by using a microscope with a stage micrometre. All measurements were performed separately for the right and left sacs. In the case where rsos.royalsocietypublishing.org R. Soc. open sci.
the embryos were at an undeveloped stage (less than or equal to 0.05 mm), their measurements were not performed.
Photographs of the shell, sacs and embryos were produced under a microscope (Leica M165 C) with an attached digital camera (Leica DFC 425, Leica Microsystems). The morphological types of the ontogenetic stages of sphaeriidae were described according to Meier-Brook [19] and Heard [35], with some additions, including the following four classes: class 1: greater than or equal to 0.05 mm, found in brood sacs which are lacking a shell or any definitive shape, usually, a cellular ball; class 2: fetal larvae: 0.2-0.4 mm, clams found in brood sacs which are lacking a shell, but having a definitive shape, such as development of the foot and visceral mass; class 3: prodissoconch larvae: 0.5-0.7 mm, clams found in brood sacs with a shell in various stages of development, but below the minimum birth size; class 4: extramarsupial larvae: greater than 0.8 mm, individuals which have broken free from the brood sacs which are fully developed. There were significant differences between the parameters of the number of embryos by size classes versus size class of parental shell and these were estimated based on the Kruskal-Wallis (multiple comparisons) test.

Structure of the population and embryonic growth
The size frequency structure of the P. casertanum population is presented in figure 3. The maximum shell length of P. casertanum in Talatinskoe Lake was 4.3 mm. The average shell length of juveniles at birth was 1.1 mm (0.8-1.6 mm) (n = 63). The proportion of juvenile (pre-reproductive) to mature bivalves was 49% to 51%, respectively, in our total sample. Among mature bivalves, the percentage of gravid molluscs was 17.5%. Furthermore, the percentage of gravid individuals with shell length 2.2-3.1 mm was 14.7% and with shell length 3.2-3.9 mm was 2.8%.
According to our data, the brood sacs with embryos in the examined population are formed when the shell length of molluscs is at least 2.2 mm. Note that the production of the brood sacs, which depends upon the location (right or left gill), does not occur simultaneously.
In some cases, the development of the brood sacs was localized at only one of the gills. The frequency histogram of sac length deviation of P. casertanum shows that a total of 69 specimens out of 95 studied (72.6% of individuals) have an asymmetric development of sacs with mean D ± s.d. = 0.104 ± 0.07 mm, min-max = 0.02−0.28 mm (figures 4-6). The mean number of embryos of one to two classes in individuals with shell-length classes of 2.4-3.1 mm was 3.6-6.2 (table 1), and with shelllength classes of 3.2-3.9 mm was 4-12.5. The mean number of embryos of three to four classes in individuals with shell-length classes 2.4-3.1 mm was 1.0-3.1, and with length of shells 3.2-3.9 mm was 3.2-7. The difference between the number of embryos of the 1-4 size classes versus shell-length class are significant (Kruskal-Wallis test: H (χ 2 ) = 14.04, p < 0.003). Variations in embryo size within individuals were also observed (figure 7). It was found that, on several occasions, the embryos of the different sizes have been found in one parental individual.

Seasonal cycle of reproduction
During the study period, the population had a high proportion of juvenile individuals (up to 50%) ( figure 8). This suggests that birth occurs between July and August and the breeding season can probably begin in July. Our results are in accordance with observations of the earlier studies [15,21,23,36,37], where P. casertanum from populations in Europe and North America presented only a single period of reproduction per year in the spring and summer months. On the other hand, in Denmark, two periods of reproduction for P. casertanum were detected in particular years, one in March-April and a second at the end of October [15]. The obtained data on the size structure of the population of P. casertanum in Lake Talatinskoe, generally, corresponds to the size parameters of this species in other parts of its distribution range [15,16]. During early August, the population of P. casertanum in Lake Talatinskoe is represented by individuals of all size classes ( figure 8). Thus, the largest individuals (3.6-4.3 mm) compose a minimum ratio. In mid-August, there was a notable mortality in the oldest classes of P. casertanum (figure 8). By the end of August, the number of gravid animals was reduced from 21 to 12%. As can be seen

Reproductive strategies
Sphaeriidae have a highly specialized reproductive system [16], being simultaneously hermaphrodites and viviparous with either synchronous or sequential brooding [9,12,35,38]. Pisidium species are synchronous brooders, developing embryos are in a brood sac that is formed by an outgrowth of the ctenidial lamellae [10,20,39]. Sphaerium and Musculium species are sequential brooders [10,39]. It is known that the Sphaeriidae show much variation in many life-history traits [16,20], such as age at first reproduction, time of egg-laying, time of embryo release, litter size, number of generations per season and others [16]. The reproductive strategies adopted by the Sphaeriidae may differ considerably  between species, geographical location and type of environment, depending on ecological factors [20,40]. During conducted comparative analyses of the seasonal dynamics of reproduction of sphaeriid clams from ephemeral and permanent ponds in Ohio (USA) [22,41], it was established that the clams which survive in the harshest environment have a higher number of reproductive strategies. In this study, in the ephemeral ponds there was prolongation of maturation time, which gave a decrease of four to five times in the number of offspring, when compared with the permanent ponds [22]. There is also a noted earlier   period of spawning embryos and shortening of duration of release for Pisidium obtusale (Lamarck 1818) in ephemeral ponds (i.e. River Volga basin, Russia) which was considered as an adaptation to the early stages of a temporary pond dessiccation process [33,42]. Simultaneously, in the population of P. casertanum inhabiting an ephemeral pond (Ontario, Canada) the litter size was larger and the generation time was shorter than that of a conspecific population in a lake [43]. In general, specific behaviour is adopted by sphaeriid bivalves under distinct stressors such as high temperature and desiccation as seen in [42,[44][45][46]. It is well known that the Arctic contains some of the most inhospitable habitat which is colonized by freshwater biota [2]. The environmental conditions are extreme for hydrobionts which inhabit those specific freshwater ecosystems, since there is a short ice-free period [1,3,47]. Consequently, for such conditions the time of brooding is reduced. It is known that temperature is the leading factor in the development of embryos [15,40]. During the short Arctic summer, there are periods of cooling when the temperature can drop to below 0 • C [27]. It has been suggested [16,48] that populations living under favourable conditions are likely to be semelparous, while populations living under unfavourable conditions are likely to be iteroparous. According to Pettinelli & Bicchierai [40], the northern populations are located in the middle Palaearctic range of the species, where iteroparous behaviour with only one litter per year seems to be a common feature. Probably, in Talatinskoe Lake, where the ice-free period is only of three months duration and the lake freezes to the bottom, to be semelparous would be irrational, however, this situation requires additional research.
According to our results, the number of embryos is correlated with the parent shell length, i.e. the greater the length of shell of the parent individuals, the higher the number of embryos (table 1 and   We established that the number of extramarsupial embryos is much lower in P. casertanum than the number of initial embryos (table 1). Many researchers, studying brooding of Pisidium species, obtained similar results [11,15,19,40]. It has been noticed in [19] that the number of eggs laid is much higher than the number of embryos which attain birth size (i.e. about half of the embryos stop growing at a length of about 0.2 mm and then die) which is a typical feature for sphaeriids. However, the specified size is  [14] and also by the present data (see the electronic supplementary material, appendix S1; table 1).
Our research also showed that embryo length within one parent individual can vary significantly (figure 7; see the electronic supplementary material, appendix S1). The obtained data agree with the observations of [19,20], where the authors found embryos of various sizes in the Pisidium species from the mountain lakes. In fact, work on several species of Pisidium described in [19,20] suggest that a chemical component, dissolved in the liquid of the brood sacs, might play a role as a growth inhibitor which is similar to those substances which effect regulation of population density in natural communities. By contrast, studies [33] on the life cycles of Pisidium species in ponds of the Volga Basin (Rybinsk Reservoir, Russia) and in the North American Great Lakes [37] found that all of the embryos were always at the same stage of development and only differed slightly in size.
Probably, in unpredictable environments with dramatic fluctuations of temperature, ephemeral ponds, drought, flood and other factors, the body size of larvae and their rates of development will have a strong influence on larval survivorship and these are precisely the characteristics that egg size most profoundly affects [48,52,53]. In addition, it is known that egg size variation is a characteristic that is susceptible to optimization by selection, in response to environmental unpredictability [52,53].
According to the theory [52,54], a female can produce eggs of very different sizes in dependents with environmental factors that increase the reproductive success [48,[52][53][54]. This mechanism was termed a 'coin-flipping' strategy, in which an individual is genetically programmed to 'flip a coin' before the spring and to choose its egg size, according to the outcome of the toss [54, p. 401]. Currently, the reproductive strategy of 'adaptive coin-flipping' has been studied on boar [48], plant parts [53], aphids [55], wasps [56] and others.
Our results indicate that release of extramarsupial embryos, presumably, occurs throughout the breeding season from July to September (i.e. before the freeze). Presumably, the laying of embryos in brood sacs occur simultaneously [10]. Then, as the embryos grow, some of them are lagging behind in their development, with respect to the others. Thus, we are likely to witness demonstrations an 'adaptive coin-flipping' of reproductive strategy. Therefore, adaptive strategies of arctic freshwater bivalves, in this case are used purposefully with respect to the support of population survival during the breeding period. In these conditions, release of embryos is not simultaneous, which is typical for representatives of the genus Pisidium [9][10][11]22,57], nor is the release occurring after a certain period of time, which would improve the chances of offspring survival.

Conclusion
The P. casertanum population in Talatinskoe Lake probably has iteroparous reproductive tactics with a single period of summer reproduction. Hence, with respect to the freshwater mollusc P. casertanum in the Arctic lake of Talatinskoe, there is a specific process of breeding, accompanied by asynchronous development and spawning of embryos. This reproductive strategy aims to improve the breeding success of the population within this extreme environment. This agrees with data from other populations with known stressors, such as in various ephemeral ponds, as well as rivers and lakes, that are exposed to extreme environmental factors. However, it is in contrast to the Pisidium species from the environmentally stable freshwater habitats of temperate latitudes which are characterized by synchronous brooding [10]. Clearly, our data fit a 'coin-flipping' strategy, where the egg size may be selected according to environmental cues [52].
Futures studies will include the different reproductive tactics of freshwater molluscs which depend upon environmental factors, including the 'coin-flipping' strategy.