I enjoyed the presented study and think it is of interest for a wide readership.
My main issues concern the sample size, the statistical analyses and how some results are interpreted. I’m aware that experiments involving breeding are time and work-intensive and come with a lot of uncertainty about the sample size (due to not all individuals breeding successfully). The small sample size and its limitations are discussed in the text, however, which is the most important aspect in my opinion.

I think the small sample size is made worse by the way the data is analysed, especially when it comes to the correlation between the number of embryos and the number of sires. It seems unnecessary to conduct independent analyses in the separate treatments (with sample sizes of only 7 and 8 litters). I don’t understand why you didn’t run a linear (or generalized linear) model to test for an effect of the number of sires on the number of embryos and included the treatment (male or female biased) as an explanatory variable. It won’t increase the statistical power, because the sample size stays equally small, but it seems the more parsimonious version compared to running three separate correlations (only male biased, only female biased and both together).
I also think the correlation between fertilized embryos and the number of males might be partly a statistical artifact, but see my specific comments below.

While this might not have been the focus of the study, it would also have been interesting to receive some more information about the males. What was the reproductive skew like and how much did it differed between the two group?

Below I list some more specific comments.

Detailed comments:

Lines 196-198: I thought the gestation time of Prairie voles is around 3 weeks. Did you find any evidence for litters born before the females were trapped again 4 weeks after the release of the male? Or would it be unlikely for the animals to mate within the first week of being released into the semi-natural enclosures?

Lines 215-217: You only mention that 15 out of 24 females bred, but it is never specified how many females bred in each of the two conditions. As far as I can see, the information is only hidden in Figure 1, and even then only when counting the raw data points . I also think how many females bred in each condition should be mentioned in the result section, instead of only here in the methods.

Lines 215 onwards: How did you deal with litters that contained only 1 pup? Were they excluded from the analyses, given that it was not possible to detect multiple paternity in those? In general, I would expect a correlation between the number of offspring and the number of sires in a litter. If a female mated with multiple males, the chance to detect it will increase with an increasing litter size. And the total number of embryos will always determine the maximal number of sires you can detect in a litter. I understand that this is something you can’t avoid of course, but I think you should at least discuss this possible complication (and maybe consider removing litters with only one pup from the analyses altogether, if you haven’t already).
Do you have an estimate about how old the embryos were? It could be interesting to see whether there is a difference in how quickly the females mated between the two contexts.

Lines 249-onwards: The first paragraph of the results section was quite difficult to read. You use several different terms to describe your two experimental conditions. You call it different contexts, different sex biased enclosures and contexts with different sex ratios. It might be easier if you would use only one of the terms to describe your two experimental conditions and use that one consistently throughout the manuscript.

Lines 260-270: I think this would be a good place to mention that you might have a higher probability to detect multiple paternity in larger litters. I don’t mean to say that this is all you observe, but I think it would be worth mentioning here.
See also my previous comments about maybe analysing everything together in one model, instead of using three separate correlations.

Lines 305-316: I don’t really see how this part is relevant to the study. It seems only tangentially related. While the section about cognitive abilities is interesting, leaving it out might make the manuscript more focused.

Lines 345-352: Is there anything known about whether the propensity to mate multiply is heritable? Could there be genetical differences between the populations?

Lines 374-onwards: Again, as I mentioned earlier. Did you really find a strong effect of multiple mating on fecundity, or was it partially driven by a higher probability to detect multiple paternity in larger litters? You can never find more sires in a litter than there are embryos, meaning large numbers of sires are only possible for large litters. And vice versa a female with only 1 pup could have mated with many sires, but you would not be able to see it.