How accurately do behavioural observations predict reproductive success in free-ranging lizards?

Behavioural ecologists often use data on patterns of male–female association to infer reproductive success of free-ranging animals. For example, a male seen with several females during the mating season is predicted to father more offspring than a male not seen with any females. We explored the putative correlation between this behaviour and actual paternity (as revealed by microsatellite data) from a long-term study on sand lizards (<i>Lacerta agilis</i>), including behavioural observations of 574 adult males and 289 adult females, and paternity assignment of more than 2500 offspring during 1998-2007. The number of males that contributed paternity to a female's clutch was correlated with the number of males seen accompanying her in the field, but not with the number of copulation scars on her body. The number of females that a male accompanied in the field predicted the number of females with whom he fathered offspring, and his annual reproductive success (number of progeny). Although behavioural data explained less than one-third of total variance in reproductive success, our analysis supports the utility of behavioural-ecology studies for predicting paternity in free-ranging reptiles.