Sampling for genetic variance in a plant nursery

We are a non-profit organization, The Vetiver Network International ([login to view URL]). We promote the use of a Vetiver grass for a wide range of environmental management applications. A common concern is that Vetiver grass could turn out to be an invasive plant, if used. Based on prior genetic characterizations (a sampling of some 250 different plants from 40 different countries), we found that about 90% of the vetivers tested came from within a narrow genetic cluster that is known to be non-fertile. We call this the "Sunshine cluster". Plants in the Sunshine cluster are propagated asexually (i.e., through clonal propagation). Recently, researchers came up with a low cost, rapid test that can detect the genetic variations that characterize the Sunshine cluster. Using this test, a plant nursery would be able to take samples, have them analyzed, and receive a lab report assessing whether or not the material they are propagating comes from from the Sunshine cluster. I am awaiting the results from a survey I sent to some 40 different, interested nurseries in order to get a feel for what the perceived variability is among their Vetiver stocks (i.e., how many different cultivars they think they may have), as well as to ascertain how many different sources they may have obtained their initial stock. With that, we will need to develop a sampling protocol for the testing. My question is on how to establish the sample size for a given nursery of population X (which may range from a few thousand to over a million); and a variability, defined as the number of different cultivars recognized in that nursery, between 1 and 10? Also, I am uncertain as to how to specify the error term in estimating sample size, as it is binary question of "Yes it demonstrates the genetic mutations consistent with being within the Sunshine Cluster" or "No it doesn't".