Samples of the seeds and seedlings of Cannabis sativa, and its dried leaves and flowerheads (marijuana), could be reliably distinguished by RAPD-PCR (Random Amplified Polymorphic DNA using the Polymerase Chain Reaction). DNA was best extracted from fresh tissues using buffers and the detergent cetyltrimethylammonium bromide; poorly dried tissue or inviable seed yielded coloured samples of degraded DNA. DNA was isolated from 51 C. sativa and two Humulus lupulus (hops) samples. Of the C. sativa samples 43 were from Australia (ten from Canberra gardens, eight from a New South Wales crop and 25 from two Queensland crops) and eight were from Papua-New Guinea (P-NG). A total of 102 different bands were obtained using four 10-nucleotide primers with arbitrarily chosen sequences. Banding patterns were compared by calculating pairwise distances using various algorithms, and presented using the neighbour-joining tree and multidimensional scaling methods. These showed a clear difference between C. sativa and H. lupulus, and separated the samples of the latter into three distinct groups; one group comprised all the P-NG samples, another the Canberra samples, and the third, the three crop samples.