Paleoclimatic data provide a highly informative if incomplete history of Arctic climate. Temperature history is especially well recorded, and it commonly allows researchers to accurately reconstruct changes and rates of changes for particular seasons. Precipitation (rain or snow) and the extent of ice on land and sea are some of the many other climate variables that have also been reconstructed. The data also provide insight into the histories of many possible causes of the climate changes and feedback processes that amplify or reduce the resulting changes. Comparing climate with possible causes allows scientists to generate and test hypotheses, and those hypotheses then become the basis for projections of future changes. Arctic data show changes on numerous time scales and indicate many causes and important feedback processes. Changes in greenhouse gases appear to have been especially important in causing climate changes (Chapter 2, section 2.4; Chapter 3, sections 3.4.1 and 3.4.4; Chapter 4, sections 4.4.1 and 4.4.2). Global climate changes have been notably amplified in the Arctic (Chapter 3, section 3.5), and warmer times have melted ice on land and sea (Chapter 6). Statistically valid confidence levels often can be attached to scientific findings, but those confidence levels commonly require many independent samples from a large population. Such a standard can be applied to paleoclimatic data in only some cases, whereas in other cases the necessary archives or interpretative tools are not available. However, expert judgment can also be used to assess confidence. The key findings here cannot all be evaluated rigorously using parametric statistics, but on the basis of assessment by the authors, all of the key findings are at least “likely” as used by the Intergovernmental Panel on Climate Change (more than 66% chance of being correct); the authors believe that the most of the findings are “very likely” (more than a 90% chance of being correct).