This study describes population genetic structure and connectivity in two sister species of Goldeneye, Barrow’s Goldeneye (Bucephala islandica) and Common Goldeneye (B. clangula) using two different markes; 7 nuclear microsatellite loci and 3,678 ddRAD-seq loci (N = 29 for Barrow’s and 32 for Common Goldeneye, mainly sampled from North America but including also a cpiåle of individuals of Common Goldeneye from Denmark (N=5 for ddRAD-seq). The authors found no genetic structure within Common or Barrow’s Goldeneye, including between North American and European samples of Common Goldeneye, in contrast with previous mitochondrial DNA, band recovery, and telemetry studies. They suggest that subadult males, may be maintaining genetic connectivity across. In addition, they identified a single F1 hybrid and estimates of gene flow <<1 migrant per generation. They conclude that either strong ecological barriers or assortative mating are likely playing a role in preventing further backcrossing and that the species have diverged some 1.6 million years before present and that the genomes of both species have been under similar evolutionary constraints.

In general, I think the authors have done much and really nice work. However, the sample size the Common Goldeneye for microsatellites is fairly small compared to the Barrow’s Goldeneye and as well as the number of loci used, which actually emphasize value of the ddRAD-seq data over the microsatellite data. The analyses are sound, interpretation well done and nicely discussed. I have only some minor comments listed below:

Line 14: There were actually only seven microsatellite loci used (not ten), weren’t there?

Line 54: Please define what is relatively long-lived, generation time?

Sampling: As you’ve marked the N used for ddRAD and microsatellites separately for the Barrow’s Goldeneye, you should perhaps do the same for the Common Goldeneye.

Much of the lab protocols could be placed into supplementary material. Now there is altogether 8 pages out of 24 of the main text describing methods.

Line 145 & 173: There is a reference to a pending manuscript for additional information of samples. I suggest to add this information also as a supplement, so that the readers don’t need to search for it from other publications.

Line 212-216: What was the reason to retain only those 8 (7) microsatellite loci out of 31?

Line 226-232: What was the purpose to sequence only 12 individuals for mtDNA control region and use samples only of the other species? Further, as the results were ambiguous due to a pseudogene, the mitochondrial parts could be deleted altogether without any effect to the purpose and results of the study.

Line 281, 295: Reference for the custom Pyhton scripts?

Line 300: Please explain why removal of contemporary hybrids ensures that obtained gene flow estimates represent North American patterns?

Line 440-484: As you suggest, the data support that immature males are the main mediators of gene flow and natal dispersal is male biased, whereas breeding dispersal not so much. Any information of dispersal distances? Telemetry/ring recovery studies of Common Goldeneyes?

Line 504: Could you add a bit more information about the pre-zygotic barriers? E.g. what kind of differences there are in habitat preferences and territorial behaviours?

Line 507: Please mention also those previous mtDNA-based estimates.

Figure 1. You could choose the colours that represent the sampling sites so that the distinction between the species would be easier. Now it is bit hard to separate them. You could also add sample sizes.

Figure 3: What is the grey individual among Barrow’s Goldeneyes in the microsatellite STRUCTURE plot (3c)?