RESEARCH NOTE Seasonal variation in gametogenesis and spawning of Mytilopsis leucophaeata, an invasive bivalve in Europe

Mytilopsis leucophaeata (Conrad, 1831) is a brackish-water species with a typically high resistance to environmental changes (Verween, Vincx & Degraer, in press). The species invaded European waters in the nineteenth century (Nyst, 1835), but it was only when it became a severe fouling species in the 1990s (Rajagopal, Van der Velde & Jenner, 1995; Verween et al., 2005), clogging European industrial cooling-water systems, that it attracted the attention of the industrial and scientific communities. Knowledge of the species’ life history is almost completely lacking, both within its original habitat and its newly invaded environment (Verween et al., in press). Populations of the closely related Dreissena polymorpha, the zebra mussel, have an annual gametogenic cycle with one or more spawning events during summer and autumn and a high degree of gametogenic synchronization (Borcherding, 1991). Temperature is a key environmental factor governing the timing of both gametogenesis and spawning in dreissenid mussels (Borcherding, 1991; Fong et al., 1995; Ram, Fong & Garton, 1996; Nichols, 1996; Claxton & Mackie, 1998) and food availability has also been suggested as an important regulator (Borcherding 1991; Nichols, 1996; Ram et al., 1996). Based on data on the presence of larval M. leucophaeata in the water column (Verween et al., 2005), we hypothesized that M. leucophaeata has only one spawning period, and does not display the bimodal pattern often found in D. polymorpha. In addition, we investigated the correlation between gametogenesis and environmental variables such as temperature, salinity and food availability. Perhaps the most useful and reliable information concerning seasonal trends in gametogenesis is obtained from histological preparations of the gonads (Seed & Suchanek, 1992). Although laborious (probably the major reason for its limited use), this method can give detailed information about the entire reproductive cycle, including the actual time of spawning. The scale of gametogenic development for both male and female mussels was determined using an original arbitrary classification system, described by Seed (1969) for M. edulis. This system has been successfully applied to D. polymorpha (Borcherding, 1991) and was compared with M. leucophaeata to distinguish more easily between the classification stages. This classification method is widely used to identify broad trends of the sexual cycle but, as in any system of arbitrary classification, intermediate stages inevitably occur, resulting in some subjectivity. To make the classification as objective as possible, Seed (1969) used multiple criteria in the assessment of each stage (Table 1). The classification stages include the resting or spent condition (0), the gamete development period (Developing I–V) and the spawning period (Spawning IV–I). The mean gonad index (MGI), defining the breeding condition of any sample, was determined monthly by multiplying the number of individuals in each stage by its numerical score (0–V, deduced from the arbitrary rating of the stage) and by dividing the sum of these products by the total number of individuals in the sample. The resulting value ranges between 0, when all the individuals are spent or resting, to V, when all individuals are sexually mature. An increase in the MGI indicates a period of development in gonadal tissue, while a decrease in the MGI indicates a period of active spawning (Seed, 1975). All data originated from an industrial site in the harbour of Antwerp, along the Schelde River in Belgium (51821.370N; 4817.300E). The study area is situated in the oligohaline zone where M. leucophaeata causes fouling problems. Mussels were collected monthly from January to December 2006. The gonads were removed carefully from the surrounding tissues, fixed in Bouin’s fluid and embedded in paraffin (608C) and sections cut at 5–10 mm were stained with toluidin blue (Pearce, 1985). An average of 25 individuals was processed every month. The slides were analysed using a Leica DMLB microscope at 200 and 400 magnification. Mussel length averaged 13.58 mm+SE 0.11. Monthly differences in gametogenic stages of M. leucophaeata were tested by analysis of variance (ANOVA). Univariate Pearson correlation matrices were used to determine correlations between environmental variables and the MGI. Gametogenesis started in January with a slowly increasing trend in late winter, accelerating through spring and early summer until the highest MGI was reached in July–August (Fig. 1), which we interpreted as the main spawning period for M. leucophaeata. Data on larval abundances of M. leucophaeata (Verween et al., 2005) endorse this pattern, with larvae appearing in the water column from June onward, with maximal densities occurring in August. The development from an egg to a D-shaped larva, large enough to be monitored in this study, takes c. 3–4 weeks (Mackie & Schloesser, 1996). In early autumn (September–October), the MGI rapidly decreased towards the winter minimum. The analysis of the gonadal conditions confirmed this pattern of gametogenesis and provided further detail. The redevelopment of the spent or resting gonads started in January, with an average of 52.0% of the individuals under developing conditions, but still 32.0% resting or spent. In the following months, the percentage of resting gonads decreased to 0% in July and August. From June onwards, the development percentages decreased, and many individuals (69.6%) began to spawn. The percentage of spawning individuals reached a maximum in September, but values exceeded 50% until December. Only a small part of the individuals were already spent or resting in September (4.2%). From October onwards, a high percentage of individuals was spent (average 26.2%+SE 6.2), with few gonads being in development (average 8.3 %+ SE 2.5). The most significant feature of the annual spawning season of M. leucophaeata was its duration; the period over which more than 50% of the individuals were spawning extended over 6 months. Thus although only a single spawning period was detected, it lasted for a very long time from June to September. This echoes the pattern observed by Bamber & Taylor (2002) Correspondence: A. Verween; e-mail: annick.verween@ugent.be