The rough periwinkle, Littorina saxatilis (Olivi, 1792), has become a model of choice for studying processes of local adaptation and ecological speciation (e.g. Johannesson et al., 2010). Interestingly, this ovoviviparous intertidal snail displays extreme female promiscuity, since a female can carry offspring sired by at least 23 males (Panova et al., 2010). Microsatellite-based parentage analysis has proved valuable for the study of multiple paternity in L. saxatilis (Makinen, Panova & Andre, 2007; Panova et al., 2010). However, three of the five markers used in these two previous studies have shown problems such as heterozygote deficiencies, null alleles and variation in allele size not equal to repeat motif (Panova et al., 2008). In many invertebrates, including molluscs in general and L. saxatilis in particular, microsatellite flanking regions contain indels and mutations causing amplification failures (null alleles), as well as cryptic repetitive DNA leading to multiple bands in polymerase chain reaction (PCR) products (McInerney et al., 2011). Such issues may significantly affect the accuracy of parentage assignment by increasing the rate of genotyping errors. To avoid these problems, the present study aimed at developing expressed sequence tag (EST)-linked microsatellites, which are likely to have more conserved flanking regions (Bouck & Vision, 2007). Six new EST-linked microsatellites were developed and combined with two recently-published loci (McInerney et al., 2009), and the performance of this new set of markers was tested for parentage analysis. EST sequences were downloaded from the Littorina Sequence Database (Canback et al., 2012), and microsatellites were identified with Tandem Repeats Finder (Benson, 1999). The 57 best di-, trior tetranucleotide microsatellites, displaying a minimum of eight repeats, a minimum of 75% matches and sufficiently long flanking regions (.40 bp), were loaded in Geneious Pro v. 5.1.7 (Drummond et al., 2010; Biomatters Ltd), and the Primer3 plugin (Rozen & Skaletsky, 2000) was used to design primer pairs with the following parameters: optimal primer length 1⁄4 20 bp (range 1⁄4 18–25 bp), optimal GC content 1⁄4 50% (range 1⁄4 30–80%), optimal Tm 1⁄4 608C (range 1⁄4 55– 618C) with a Tm difference not larger than 28C between the forward and reverse primers from the same pair and product length 1⁄4 100–400 bp. Finally, 36 primer pairs were selected for further PCR amplification tests. A ‘universal tail’ was added to the 50-end of each forward primer to reduce the genotyping cost (Schuelke, 2000). Three distinct ‘universal tails’ were used (tail1, tail2 or tail4), as described by Real, Schmidt & Hughes (2009). A PIG-tail (50-GTTTCTT) was also added to the 50-end of each reverse primer to avoid ‘plus-A’ PCR artefacts (Brownstein et al., 1996). In addition, three microsatellites developed by McInerney et al. (2009), with perfect trinucleotide (Lsax18 and Lsax20) or tetranucleotide repeats (Lsax13), were included in the study. All 39 microsatellites were tested with a population sample of L. saxatilis collected on the western Swedish coast, on the island of Salto (588530N, 118100E). Individuals were stored in ethanol at 2208C. DNA was extracted using the EZNA Mollusc DNA Extraction kit following the manufacturer’s instructions. Amplification tests were performed on eight individuals for all 39 loci. Each EST-linked microsatellite was amplified by PCR in a 10-ml volume, containing 1 RBC reaction buffer (RBC Bioscience), 1.0–2.0 mM MgCl2, 0.1 mM each dNTP, 0.125 mM ‘universal’ primer end-labelled with a WellRED dye (D2, D3 or D4; Sigma-Aldrich), 0.0125 mM tailed forward primer, 0.125 mM reverse primer, 0.025 U ml RBC Taq DNA polymerase (RBC Bioscience) and 1 ml DNA template. The three microsatellites published by McInerney et al. (2009) were amplified following the same protocol as above but with 0.15 mM forward primer end-labelled with a WellRED dye (Sigma-Aldrich) and 0.15 mM reverse primer. PCR amplifications were performed in a Veriti Thermal Cycler (Life Technologies). A touchdown procedure was incorporated in the thermal cycling regime to increase the stringency of the PCRs: 948C for 3 min, Ta1 for 2 min, 728C for 1 min, (948C for 30 s, Ta1 218C for 30 s [218C per cycle until Ta2], 728C for 1 min) 9–12 cycles, (948C for 30 s, Ta2 for 30 s, 728C for 1 min) 22–30 cycles, 728C for 5 min. A total of 35 PCR cycles were run by default, but this was increased to 40 cycles when the peak intensity was too low. Details of the PCR protocol, i.e. MgCl2 concentration, number of PCR cycles and type of ‘universal’ tail, are provided in Table 1 for the successfully amplified loci. Amplification success was checked with electrophoresis on 1% agarose gel in 0.5 tris-borate-EDTA. Finally, PCR products were run on a Beckman Coulter CEQ 8000 capillary sequencer