The American Community Survey (ACS) produced a series of three- and 5-year period estimates for 34 ACS counties using data from 1999-2005. This series of period estimates provided a large amount of information regarding the variance distributions for multiyear ACS estimates for the first time. The focus of this paper is on answering three questions: how much a new weighting methodology reduced variances at the place and census tract levels, what are the variance characteristics of estimates at the block group level (the smallest geographic level published in the ACS five-year products), and the relationship between the ACS tract level variances and corresponding variances from the Census 2000 Long Form.