The increasing popularity of cellular telephones as the primary communication medium between individuals is ushering in a new era for telephone surveys. However, most of what we know about cellular telephone ownership and use in the United States has come from unsystematic or incomplete sources. No one previously has attempted to assess the feasibility of conducting a survey that employs a cellular number frame. With the spread of payment plans that eliminate some of the charges for incoming cellular calls, such a survey no longer seemed impossible to administer. Concurrently, the decision by the Federal Communications Commission to allow owners to switch from provider to provider or from landline to cellular telephones without changing their number created a compelling need for such an effort since even the standard RDD sample would contain cellular telephone numbers after November 2003. In this paper I report the results of a national survey conducted in 2003 using a cellular number frame. This survey indicates what dilemmas sampling statisticians, survey methodologists, and interviewers alike will face in the twenty-first century. Although the administration of this kind of survey differs in many ways from the administration of a conventional RDD survey, I will focus here on the issue that seems to be of crucial importance—bias both from noncoverage and nonresponse. It is clear that at the present time a survey based only on cellular numbers is not representative of the general adult population in the United States. It is also probably not representative of cellular owners and users. The proposed analysis will compare demographic characteristics and selected behaviors and attitudes reported by respondents to this survey with similar measures from a companion RDD survey  conducted at approximately the same time using identical methods. In order to detect both noncoverage and nonresponse bias, the results from both of these surveys will be compared to estimates derived from the 2000 United States Census, the March 2003 Current Population Survey, and the 2002 General Social Survey.