Neonatal reproductive failure should occur when energetic costs of parental investment outweigh fitness benefits. However, little is known about the drivers of neonatal reproductive failure in free ranging species experiencing continuous natural variation in predator abundance and in the energetic and fitness costs and benefits associated with parental investment. Long-term comprehensive studies are required to better understand how biotic, abiotic, and life history conditions interact to drive occurrences of reproductive failure in the wild. Using 24 years (1987-2011) of reproductive data from a northern boreal population of North American red squirrels in southwestern Yukon, we examined the effects of predator abundance, energetics (resource availability, ambient temperature and litter size), and fitness benefits (probability of overwinter juvenile survival and maternal age) on occurrences of neonatal reproductive failure (494/2670 reproductive attempts; 18.5%). Neonatal reproductive failure was driven by a combination of predator abundance, and the energetic and fitness costs and benefits of parental investment. The abundance of mustelids and maternal age were positively related to the occurrence of neonatal reproductive failure. High energy costs associated with a combination of low resource availability and cold ambient temperatures or large litters, corresponded to increased occurrences of neonatal reproductive failure. However, the strength of these relationships was influenced by variation in juvenile overwinter survival (i.e. fitness benefits). We provide evidence that predation pressure is an important driver of neonatal reproductive failure. In addition, we found a trade-off occurs between resource-dependent energetic and fitness costs and benefits of raising the current litter to independence

Reproductive failure of red squirrels and associated measurementsData collected in the field from 1989 to 2012 on reproductive failure in red squirrels and associated variables. Columns are: LitID- litter ID # in analysis, Grid - study area in which squirrel was living, YR - Year of reproductive attempt, LN - reproductive attempt number of that year for the female (all are the first litter), BDATE - estimated parturition date of litter (based on live trap monitoring of females), FoodSup - whether squirrels were provided food supplementation (1) or not (0), MOMID - unique ear tag numbers of female squirrels, litsize - number of pups in litter, PCI - previous years spruce cone index (log transformed), weas - annual average number of weasel tracks/100km counted in area, marten - annual average number of marten tracks/100km counted in area, pYRSurv - proportion of that years cohort of juveniles that survived the winter to the following year, cold - coldest 4-day average temperature experienced during the first 14 days of lactation (Environment Canada’s National Climate Data and Information Archive (www.climate.weatheroffice.gc.ca)), mBY - birth year of the female, pred - annual average number of mustelid (marten and weasel) tracks/100km counted in area, MAge - maternal age.reproductive_failure.csv