G success (binomial distribution), and burrow was added as an supplementary random effect (because a few of the tracked birds formed breeding pairs). All means expressed in the text are ?SE. Data were log- or square root-transformed to meet parametric assumptions when necessary.Phenology and breeding successIncubation lasts 44 days (Harris and Wanless 2011) and is shared by parents alternating shifts. Because of the difficulty of intensive direct observation in this subterranean nesting, easily disturbed species, we estimated laying date indirectly using saltwater immersion data to detect the start of incubation (see Supplementary Material for details). The accuracy of this method was verified using a subset of 5 nests that were checked daily with a burrowscope (Sextant Technology Ltd.) in 2012?013 to determine precise laying date; its accuracy was ?1.8 days. We calculated the birds’ postmigration laying date for 89 of the 111 tracks in our data set. To avoid disturbance, most nests were not checked directly during the 6-week chick-rearing period following incubation, except after 2012 when a burrowscope was available. s11606-015-3271-0 Therefore, we used a proxy for breeding success: The ability to hatch a chick and rear it for at least 15 days (mortality is highest during the first few weeks; Harris and Wanless 2011), estimated by direct observations of the parents bringing food to their chick (see Supplementary Material for details). We observed burrows at dawn or dusk when adults can frequently be seen carrying fish to their burrows for their chick. Burrows were deemed successful if parents were seen provisioning on at least 2 occasions and at least 15 days apart (this is the lower threshold used in the current method for this colony; Perrins et al. 2014). In the majority of cases, birds could be observed bringing food to their chick for longer periods. Combining the use of a burrowscope from 2012 and this method for previous years, weRESULTS ImpactNo immediate nest desertion was witnessed posthandling. Forty-five out of 54 tracked birds were recaptured in following seasons. OfBehavioral Ecology(a) local(b) local + MediterraneanJuly August September October NovemberDecember January February March500 km (d) Atlantic + Mediterranean500 j.neuron.2016.04.018 km(c) Atlantic500 km500 kmFigure 1 Example of each type of migration routes. Each point is a daily GSK089 position. Each color represents a different month. The colony is represented with a star, the -20?meridian that was used as a threshold between “local” and “Atlantic” routes is represented with a dashed line. The breeding season (April to mid-July) is not represented. The points on land are due to low resolution of the data ( 185 km) rather than actual positions on land. (a) Local (n = 47), (b) local + EW-7197 web Mediterranean (n = 3), (c) Atlantic (n = 45), and (d) Atlantic + Mediterranean (n = 16).the 9 birds not recaptured, all but 1 were present at the colony in at least 1 subsequent year (most were breeding but evaded recapture), giving a minimum postdeployment overwinter survival rate of 98 . The average annual survival rate of manipulated birds was 89 and their average breeding success 83 , similar to numbers obtained from control birds on the colony (see Supplementary Table S1 for details, Perrins et al. 2008?014).2 logLik = 30.87, AIC = -59.7, 1 = 61.7, P < 0.001). In other words, puffin routes were more similar to their own routes in other years, than to routes from other birds that year.Similarity in timings within rout.G success (binomial distribution), and burrow was added as an supplementary random effect (because a few of the tracked birds formed breeding pairs). All means expressed in the text are ?SE. Data were log- or square root-transformed to meet parametric assumptions when necessary.Phenology and breeding successIncubation lasts 44 days (Harris and Wanless 2011) and is shared by parents alternating shifts. Because of the difficulty of intensive direct observation in this subterranean nesting, easily disturbed species, we estimated laying date indirectly using saltwater immersion data to detect the start of incubation (see Supplementary Material for details). The accuracy of this method was verified using a subset of 5 nests that were checked daily with a burrowscope (Sextant Technology Ltd.) in 2012?013 to determine precise laying date; its accuracy was ?1.8 days. We calculated the birds' postmigration laying date for 89 of the 111 tracks in our data set. To avoid disturbance, most nests were not checked directly during the 6-week chick-rearing period following incubation, except after 2012 when a burrowscope was available. s11606-015-3271-0 Therefore, we used a proxy for breeding success: The ability to hatch a chick and rear it for at least 15 days (mortality is highest during the first few weeks; Harris and Wanless 2011), estimated by direct observations of the parents bringing food to their chick (see Supplementary Material for details). We observed burrows at dawn or dusk when adults can frequently be seen carrying fish to their burrows for their chick. Burrows were deemed successful if parents were seen provisioning on at least 2 occasions and at least 15 days apart (this is the lower threshold used in the current method for this colony; Perrins et al. 2014). In the majority of cases, birds could be observed bringing food to their chick for longer periods. Combining the use of a burrowscope from 2012 and this method for previous years, weRESULTS ImpactNo immediate nest desertion was witnessed posthandling. Forty-five out of 54 tracked birds were recaptured in following seasons. OfBehavioral Ecology(a) local(b) local + MediterraneanJuly August September October NovemberDecember January February March500 km (d) Atlantic + Mediterranean500 j.neuron.2016.04.018 km(c) Atlantic500 km500 kmFigure 1 Example of each type of migration routes. Each point is a daily position. Each color represents a different month. The colony is represented with a star, the -20?meridian that was used as a threshold between “local” and “Atlantic” routes is represented with a dashed line. The breeding season (April to mid-July) is not represented. The points on land are due to low resolution of the data ( 185 km) rather than actual positions on land. (a) Local (n = 47), (b) local + Mediterranean (n = 3), (c) Atlantic (n = 45), and (d) Atlantic + Mediterranean (n = 16).the 9 birds not recaptured, all but 1 were present at the colony in at least 1 subsequent year (most were breeding but evaded recapture), giving a minimum postdeployment overwinter survival rate of 98 . The average annual survival rate of manipulated birds was 89 and their average breeding success 83 , similar to numbers obtained from control birds on the colony (see Supplementary Table S1 for details, Perrins et al. 2008?014).2 logLik = 30.87, AIC = -59.7, 1 = 61.7, P < 0.001). In other words, puffin routes were more similar to their own routes in other years, than to routes from other birds that year.Similarity in timings within rout.