A program of estimates covering eight classes of nursery products and four selected out flowers grown in 5 States was undertaken in late 1956 by the Fruit and Vegetable Statistics Branch of the Agricultural Estimates Division, AMS. The initial survey was an almost complete enumeration with estimates being required for only a few small nonrespondents whose size of operation had previously been established. To reduce time and costs, random sampling was tested in a 1957 survey in these same States. A sample was drawn to measure 1957 sales in relation to those made in 1956 with sampling errors of about 2 percent by States for each item. The sample was checked against (1) the results obtained from a general mailing to all producers with several followups and (2) the results obtained from similar mailings and followups of large nonrespondents by personal interview. Final results show that more accurate estimates are obtained from general mailings to all known producers with selective followups of the larger nonrespondents.