The fundamental long-term objective of the seabird component of the Palmer LTER (PAL) has been to identify and understand the mechanistic processes that regulate the mean fitness (population growth rate) of regional penguin populations. Two hypotheses have guided this research, with one suggesting that population mean fitness is best explained by changes in regional krill biomass, and the other proposing that long-term changes in sea ice affects mean fitness by tipping the balance in favor of one species over another in accordance with species-specific evolved life history affinities to sea ice. Although these hypotheses are not mutually exclusive, current evidence in the PAL region tends to favor the latter over the former. Since the inception of PAL, Adélie penguin populations have effectively collapsed, while those of gentoo and chinstrap penguins have increased dramatically, trends that are spatially and temporally coherent with decreasing regional sea ice duration. Adélie penguins are an ice-obligate polar species whose life history is intimately linked to the presence of sea ice, while chinstrap and gentoo penguins are ice-intolerant species whose life histories evolved in the sub-Antarctic, where sea ice is a less permanent feature of the marine ecosystem. In contrast, although krill constitute the most important component of the summer diets by mass of these three penguin species, changes in PAL krill abundances have exhibited no long-term trends, and thus fail to explain the divergent patterns in penguin populations evident in our time series. The arrival chronology of adult Adélie penguins on Humble Island is documented annually through island-wide censuses performed as ice and weather conditions permit. Recorded data (numbers of adults present) provide a measure of the number of adults arriving daily at the breeding colonies, a metric that is sensitive to environmental conditions such as sea ice extent during late winter and early spring. These data are also used in combination with other metrics to determine the optimal window for other, more extensive area-wide breeding population censuses (see CENSUS).