It has long been asserted that increased gambling availability leads to increased participation and harm (availability hypothesis). It has also been proposed that over time participation and harm decrease even when availability continues to rise (adaptation hypothesis). New Zealand has national gambling and problem gambling surveys dating from the mid-1980s. They include five-yearly surveys of gambling behaviour and attitudes from 1985 to 2005 and surveys of gambling and problem gambling in 1990 and 1999. The National Gambling Study (NGS) (2012–2015) was in part designed to assess changes in behaviour, attitudes and gambling-related harm since the 1999 and 2005 surveys. Selected NGS data are examined in relation to data from earlier surveys. A national lottery, instant lotteries and electronic gaming machines were introduced during the late 1980s. Since then, gambling availability continued to increase. Participation in most newly introduced gambling forms increased markedly, usually within the first year or two, and then decreased, often substantially. Following an initial rise, contrary to the availability hypothesis, overall participation declined. This was accompanied by a decline in problem gambling prevalence. These findings are consistent with adaptation. Since 2000, while participation continued to fall, problem gambling prevalence plateaued. This finding appears to be at variance with both the availability and adaptation hypotheses. It points to the importance of factors other than gambling availability in determining problem gambling and related harm. Possible reasons for the plateauing of problem gambling in the face of substantial reductions in participation are considered. They have implications for future research and policy.