the role of male quality in sequential mate choice: pregnancy replacement in small mammals?

Females mainly increase their reproductive success by improving the quality of their mates and need to be discriminative in their mate choices. Here we investigate whether female mammals can trade up sire quality in sequential mate choice during already progressed pregnancies. A male-induced pregnancy termination (functional ‘Bruce effect’) could thus have an adaptive function in mate choice as a functional part of a pregnancy replacement. We used bank voles (Myodes glareolus) as a model system and exchanged the breeding male in the early second trimester of a potential pregnancy. Male quality was determined using urine marking values (UMV).  Females were offered a sequence of either high then low-quality male (HL), or a low then high-quality male (LH). The majority of females bred with the high-quality male independent of their position in the sequence, which may indicate a pregnancy replacement in LH but not in HL. The body size of the second male, which could have been related to the coercion of females by males into re-mating, did not explain late pregnancies. Thus, pregnancy replacement, often discussed as a counterstrategy to infanticide, may constitute adaptive mate choice in female mammals.

Methods
Each female was paired with male 1 for a duration of 7 days (5-8 days in dataset 2), after which the male was removed and replaced with the second male (M2), which stayed with the female for another 7 days. For the experiment males were placed into females' cages to mimic the natural situation where males would enter females' territory. On an experimental day 14, pairs were separated and all animals were transferred back into their single-home cages. Due to the bank vole pregnancy lasting 20 ± 2 days (mean ± SD, we expected litters born on experimental days 18 to 32. Cages were controlled daily for births.

We challenged our paternity assignment for the litters born in the second peak (days 25 – 29) in dataset 1, using genetic paternity analyses. We assigned paternity to 10 out of 11 families using genetic parentage analyses and confirmed 9 out of 10 of our assignments (90%), and one litter had a change in the assignment. Multiple males sired none of the litters. One family did not have any of the males confirmed as fathers, due to missing genetic samples.