Our analyses reveal that despite short geographic distances between several of the breeding colonies of Nazca boobies, there is substantial genetic differentiation within the Galapagos archipelago and that three genetically distinct populations occur within the archipelago, based on the Bayesian clustering analysis. In contrast, very weak to no population genetic structure was found in the great frigatebird using both mitochondrial and nuclear markers and we found evidence of migration of individuals between multiple colony pairs. Most of the migration rates calculated for Nazca boobies were low, with the exception of high levels of nearly unidirectional gene flow were detected between two Nazca booby colonies, Española and Genovesa. We found that several of the larger migration rates (large relative to the general trend of low numbers of individuals moving between most colony pairs) were from Española to other colonies, indicating that it might be a source population. The pronounced genetic differentiation in Galapagos Nazca boobies detected here corroborates previous mark-recapture studies that demonstrated very limited natal and breeding dispersal of Galapagos Nazca boobies (Huyvaert and Anderson 2004).
Diversity within populations
Genetic diversity estimates within each population and across all populations were reasonably high and even for both species across populations. Our estimate of 58% (Nazca booby) and 65% (great frigatebird) heterozygosity is similar to values reported for other Galapagos taxa such as the Galapagos dove (56–65%) (Santiago-Alarcon et al. 2006) and the flightless cormorant (51–66%) (Duffie et al. 2009) and higher than Galapagos penguins (44%) (Nims et al. 2008) and Galapagos mockingbirds (Mimus spp.) (35%) (Hoeck et al. 2010). The caveat when comparing genetic diversity calculated from microsatellites between studies is that ascertainment bias can result from investigators selecting for polymorphic loci during primer development (Ellegren et al. 1995). Additionally, when microsatellites are used for species other than the one they were designed for (as is our case for Nazca boobies but not great frigatebirds), this ascertainment bias can lead to artificial differences due to lower polymorphism in the non-focal species (Brandström and Ellegren 2008). Whenever possible, we selected markers with a medium and comparable number of alleles. Nevertheless, results should be interpreted with caution due to ascertainment bias and the fact that population size influences genetic diversity.
Evidence for recent bottlenecks was detected in the Española, Genovesa, and San Cristobal colonies of Nazca boobies. This could be due to the El Niño Southern Oscillation (ENSO) events that raise sea surface temperature, which can negatively affect marine life in Galapagos. The 1986–1987 ENSO event, while less severe than the one in 1982–1983, caused Nazca boobies to either suspend breeding or adjust the timing of their breeding cycle (Anderson 1989). These results must be interpreted with caution due to recent literature review demonstrating underestimation of bottlenecks by moment-based estimators like the one used in this analysis (Peery et al. 2012). Additionally, Peery et al. (2012) point out that the proportion of multistep mutations is often underestimated in microsatellite datasets and therefore bottleneck tests can spuriously detect bottlenecks in stable populations.
Haplotype diversities estimated from mtDNA were fairly high (h = 0.886 for Nazca boobies, 0.644 for great frigatebirds), compared with the Galapagos flycatcher (Myiarchus magnirostris) (h = 0.491) (Sari and Parker 2012) and comparable to the Galapagos hawk (Buteo galapagoensis) (h = 0.671) (Bollmer et al. 2006). Four island colonies of Nazca boobies had three or more unique mtDNA haplotypes (Darwin = 3, Genovesa = 3 San Cristobal = 4, Wolf = 4), and the most genetically distinct island was Darwin. For great frigatebirds, 64 of 108 individuals shared one common haplotype. All island populations except Darwin had at least two unique mtDNA haplotypes and there were four haplotypes that were shared between N. Seymour and Española. The star-like shape of the frigatebird haplotype network with one common haplotype and several unique haplotypes differing by one mutational step resembles that of an organism undergoing a demographic expansion. However, we did not detect a significantly negative value for Tajima's D (−1.64, close to significant but P > 0.05).
Differentiation between populations
As predicted, population differentiation was more pronounced among Nazca booby populations compared with populations of great frigatebirds, most likely driven by differences in degree of natal and breeding philopatry. According to the multilocus dataset, great Frigatebirds showed very weak to no genetic structure, with the largest FST, 0.0396, between Darwin and Wolf, the two islands closest in proximity. Even with the Locprior setting in STRUCTURE, we detected no population subdivision among great frigatebird colonies (Fig. 3, shown at k = 3 for comparison with the three groups detected in the Nazca booby). Interestingly, the mitochondrial genes provide evidence for weak differentiation between N. Seymour and Darwin and N. Seymour and Wolf. These discrepancies in overall pattern calculated using different markers could be due to the timescale on which the markers provide the best resolution (mtDNA most useful for a more historical perspective and microsatellite DNA best for more recent estimates of divergence) and/or any sex bias in dispersal (maternally inherited mtDNA vs. biparentally inherited nuclear microsatellites). Although we have evidence that Galapagos great frigatebirds are genetically distinct from their non-Galapagos conspecifics (F. Hailer, unpubl. data), the birds breeding within the archipelago appear to be exchanging genes at a rate that swamps any effects of philopatry. The archipelago-wide average migration rate for great frigatebirds is similar to the one calculated for Nazca boobies; however, individual rates between colonies are more variable in the frigatebird while Nazca booby migration rates are all quite low aside from substantial movement of individuals from Española to Genovesa and from Wolf to Darwin. There is historical evidence from great frigatebirds breeding in the northwestern Hawaiian Islands demonstrating that the locations of breeding colonies are dynamic and the patterns may be explained by changes in vegetation important for nesting (Cohen and Dearborn 2004). Within Galapagos, we have anecdotal evidence of similar dynamics; the breeding colony sampled on Española in 2007 was not present in 2010 and there were no indications of prior recent breeding (old nests, chicks, juveniles). Aside from lower natal and breeding philopatry, another explanatory factor could be lack of philopatry to non-breeding site. Friesen et al. (2007) found philopatry to non-breeding site to be a strong predictor of population genetic structure. We do not have good information on philopatry during the non-breeding season for this population, but it has been shown that great frigatebirds travel great distances during the non-breeding season, and therefore may contribute to lower philopatry during the non-breeding season (Weimerskirch et al. 2006). At-sea distribution may also play a role in shaping population structure. Telemetry studies of great frigatebirds in the northwestern Hawaiian Islands indicate that, while there is a lot of variation, most of the foraging trips of great frigatebirds caring for chicks are within 200 km from their nesting colony (Gilmour et al. 2012).
Nazca boobies showed pronounced genetic differentiation. As predicted, population differentiation, as measured by FST calculated with the multilocus dataset, was statistically significant between all but two Nazca booby population pairs (Genovesa-Española; Darwin-Wolf). The gene flow between Darwin and Wolf is not surprising given that they are separated by only 38 km. Gene flow between Genovesa and Española, separated by 194 km, but not between San Cristobal and either Genovesa (140 km) or Española (87 km) is a bit more puzzling. The western tip of San Cristobal is slightly east of a straight line between Española and Genovesa, but the main seabird colonies are located on the extreme northeastern tip of the island, also the most eastern point in the archipelago, with other smaller colonies along islets on the north side. Española birds dispersing in a north-northwestern direction, and therefore not passing over the colony on San Cristobal, would explain our estimates of archipelago-wide directional migration rates, and suggest that most gene flow occurs in a northern or northwestern direction. Interestingly, gene flow was also highest between Galapagos doves sampled on Española and Genovesa, although San Cristobal was omitted from the analyses due to small sample size (Santiago-Alarcon et al. 2006). Similarly, Arbogast et al. (2006) found that Galapagos mockingbirds from Española, Genovesa, and San Cristobal had very similar mtDNA despite being considered different species.
Mitochondrial φST values for Nazca boobies showed a different pattern than FST's calculated with microsatellites. φST values were low for most colony pairs except for Darwin and all other colonies, which showed high levels of differentiation. Again, these discrepancies are likely due to the differences in the strength of the particular marker in resolving divergence at different timescales and could also reflect sex differences in dispersal as mtDNA provides only maternally inherited information. Although genetic distinctiveness of Darwin birds was not seen in the microsatellite analysis of pair-wise differentiation, we find a similar pattern of differentiation in the extreme corners of the archipelago: Darwin is the most northern and most western of the islands while San Cristobal is the most eastern island currently above sea level. This pattern is evident in Darwin's finches (Geospiza, Camarhynchus, Catospiza, and Certhidea spp.), where peripheral populations were found to be more genetically distinct (Petren et al. 2005). However, the larger Nazca booby colonies we sampled for this study are all arguably peripheral, so it is difficult for us to provide much support for the claim that peripheral isolation is driving this pattern of population differentiation in our system. Finally, despite the fact that, depending on the molecular markers used, different colonies emerge as the most genetically distinct, there are consistencies between the mtDNA and the multilocus datasets; several of the pair-wise relationships tell the same story for both marker types (e.g., Española and Genovesa, Darwin and San Cristobal, Darwin and Española).
No strong relationship was found between geographic distance and genetic differentiation of Nazca boobies using either mtDNA or microsatellite data. A Mantel test did detect a significant isolation by distance relationship using φST, but it appeared to be an artifact of a few points, only explaining 14% of the variation in the data. A positive relationship between geographic distance and genetic differentiation was found in Galapagos passerine birds (Petren et al. 2005; Hoeck et al. 2010), Galapagos hawks (Bollmer et al. 2005), and in flightless cormorants (Duffie et al. 2009) where distance-limited dispersal is not surprising; however, it is not surprising that we do not find isolation by distance effects in a vagile seabird on such a small geographic scale.
The Bayesian clustering analysis detected three distinct populations of Galapagos Nazca boobies: San Cristobal, Genovesa and Española, and Darwin and Wolf. These results are consistent with the genetic uniqueness of San Cristobal birds (this population, along with Genovesa, had the greatest number of private alleles), and the relative isolation of Darwin and Wolf compared with any other islands in the archipelago. Migration rate estimates indicate that the highest level of gene flow occurs from Española to Genovesa and from Wolf to Darwin. Interestingly, there is negligible gene flow from Genovesa to Española. The majority of migration rate estimates >0.01 are in a north or northwestern direction, the direction of the prevailing winds.
Galapagos Nazca booby colonies are strongly genetically structured, especially when considering the small geographic scale while great frigatebirds are not. Regarding the structure detected in the Nazca booby, some Sulidae species show strong phylogeographic signals and/or population genetic structure (e.g., brown booby (Morris-Pocock et al. 2011); red-footed booby (Morris-Pocock et al. 2010b), while others do not (e.g., blue-footed booby (Taylor et al. 2011); Peruvian booby (Taylor et al. 2010b)). A possible explanation for the lack of structure in the blue-footed and Peruvian booby populations is their specialization to cold-water upwelling environments such as the Humboldt Current system. When ENSO events disrupt the upwelling, successful reproduction and survival could depend on movement of individuals to more suitable breeding colonies (Taylor et al. 2011). Population differentiation in the Galapagos Nazca booby and other Sulidae is most likely due to strong natal philopatry. Median natal dispersal distances for Española Nazca boobies were 105 m for females and 26 m for males (Huyvaert and Anderson 2004). Only one of 198 breeding dispersal distances (breeding sites between years) within the Punta Cevallos, Española colony was greater than 25 m (Huyvaert and Anderson 2004). Documented natal dispersal from Española to other Nazca booby colonies was rare, with an estimate of 1.3% of banded nestlings moving to other surveyed islands (excluding Darwin and Wolf) (Huyvaert and Anderson 2004). This value is lower than our estimated mean migration rate across the archipelago, 0.037, but that is not surprising given that mark-recapture techniques are sure to miss some natal dispersal events leading to an underestimate. Seventeen band records were reported outside of Galapagos, indicating Galapagos Nazca boobies can disperse long distances, but will only do so rarely (Huyvaert and Anderson 2004). We gain some insights into at-sea distribution from recent telemetry work on Española Nazca boobies rearing chicks (Zavalaga et al. 2012). Individuals tended to stay within 200 km from their nesting colony during single-day trips and a maximum of 329 km were recorded for longer foraging trips (Zavalaga et al. 2012). These data, and our findings, clearly illustrate what has been called “the seabird paradox” (Milot et al. 2008) where some pelagic species show strong population genetic differentiation despite being highly mobile (Friesen et al. 2007). This paradox raises important questions involving natal and breeding dispersal, benefits of philopatry and coloniality, potential barriers (physical and non-physical) to dispersal, and colony persistence that are fundamental to our understanding of evolution in seabirds.