DATA1_n=49.csvData for all primary analyses using n = 49 males. Includes standardized variables labeled as '...cen', along with rank transformed variables '...Ran'. Relative reproductive success (relRS) Also includes raw variables and body condition estimates for each male. HRA = home range area, SVL = snout-vent length. Move/pro.move = proportion of time moving; days.active/Totday = total days active.DATA1_nDATA2_n=56Data for all secondary analyses using n = 56 males. Note that some home range re-sightings are below 8 and are considered to possibly underestimate home range area. Includes standardized variables labeled as '...cen', along with rank transformed variables '...Ran'. Relative reproductive success (relRS) Also includes raw variables and body condition estimates for each male. HRA = home range area, SVL = snout-vent length. Move/pro.move = proportion of time moving; days.active/Totday = total days active.DATA2_nFULLHRADATAData for analyses involving all individuals with complete home range data n = 66 males. Note that this does not include data on movement rates, which may be important. The analysis was simply used to test whether estimates were over-inflated as a result of the reduced sample size where we had accurate and complete data (i.e. n = 49). Includes standardized variables labeled as '...cen'. Relative reproductive success (relRS) Also includes raw variables and body condition estimates for each male. HRA = home range area, SVL = snout-vent length. Move/pro.move = proportion of time moving; days.active/Totday = total days active.

Alternative reproductive tactics (ARTs) are predicted to be the result of disruptive correlational selection on suites of morphological, physiological and behavioral traits. ARTs are most obvious when they occur in discrete morphs with concomitant behavioral tactics. However, ARTs driven by behavior, in species lacking obvious phenotypic differences, are rarely documented and poorly understood. We quantified selection acting on phenotypic traits predicted to characterize ARTs by observing marked lizards in six semi-natural populations. We quantified reproductive fitness for each male using 6 microsatellite DNA loci from 226 offspring born to 56 females. Candidate models containing directional and correlational selection gradients were equally supported. As predicted, large males with large home ranges and large males who were observed frequently had the highest reproductive success. We also found evidence that large males that moved little, but were observed frequently and large males which moved frequently, but were observed little, were predicted to have high fitness. Model predictions support our verbal hypothesis regarding the phenotypes characterizing ARTs and suggest that large males may be adopting subtly different tactics to acquire paternity. Our results suggest that disruptive correlational selection between behavioral traits may drive the evolution of ARTs in 'cryptic' systems that lack overt polymorphisms.