(1998). A fragment of approximately 700 bp from the 5′ end of the maternally-inherited mtDNA control region was amplified and sequenced according to Hamner et al. (2012). Sequences were aligned and edited using Geneious Pro

v5.5.2 (BioMatters). Haplotypes were initially assigned based on the 360 bp reference sequences of the 22 haplotypes previously identified for Hector’s and Maui’s dolphins (Pichler et al. 1998, Pichler and Baker 2000, Pichler 2002, Hamner et al. 2012), however several of these haplotypes were further resolved based on alignment with longer 576 bp sequences. All samples were genotyped for 21 microsatellite loci using published cetacean Tanespimycin cost primers (Table 1). For the “SGUI” loci and TtruGT48, each 10 μL PCR reaction contained 1 ×  PCR II buffer, 2.5 mM MgCl2, 0.04 μM of the forward primer with M13 tag, 0.4 μM reverse primer, 0.4 μM fluorescent label with M13 tag, 0.2 mM dNTP, 20 mg/mL bovine serum albumin (BSA), 0.25 units p38 MAPK inhibitor Platinum Taq (Invitrogen) and 10–20 ng/μL DNA template, and were amplified using the thermocycling profile of Cunha and Watts (2007) with modifications to the annealing temperature specified in Table 1. For all other loci, each 10 μL PCR reaction contained 1 ×  PCR II buffer, 2.5 mM MgCl2, 0.4 μM each primer, 0.2 mM dNTP, 0.125 units Platinum Taq (Invitrogen) and 10–20 ng/μL DNA template, and were

amplified using the following thermocycling profile: 93°C for 2 min; (92°C for 30 s, TA for 45 s, 72°C for 50 s) × 15; (89°C for 30 s, TA for 45 s, 72°C for 50 s) × 20; 72°C for 3 min, with the annealing temperatures check details (TA) stated

in Table 1. Products were run on an ABI 3130XL DNA Analyzer and allele peaks were binned and visually verified using GENEMAPPER v.3.7 (Applied Biosystems). To minimize genotyping error, each amplification and sizing run included a negative control to detect contamination and 10 internal control samples to ensure comparable allele sizing across all runs and to estimate genotyping error. A genotyping error rate was calculated by dividing the number of incongruent allele calls by the total number of alleles compared for the samples that were genotyped twice (Bonin et al. 2004). Genotypes were compared to identify replicate samples of the same individual using CERVUS v. 3.0 (Kalinowski et al. 2007). The probability of identity (P(ID)) and probability of identity for siblings (P(ID)sib) for each locus and across all loci were calculated in GenAlEx v. 6.1 (Peakall and Smouse 2006). To avoid false exclusion, initial matching allowed for up to five mismatching loci, and we examined each of these “relaxed matches” for potential allelic dropout or processing error, and repeated them as needed for confirmation. Sex and mtDNA haplotypes were subsequently compared to support our confidence in correctly identifying replicate samples.