Two experiments were conducted at Giza and Sids Research Stations in 1999%2000 season, to determine the optimum plant sample size for yield and its components and morphological characters in lentil. Five plant sample sizes (G, 20, 30, 40 and 50 plants/plot) with the whole plant-plot population were used with the new lentil varieties Giza 4 and Giza 51. Results indicated that the characters, which had high variation, needed large sample size. The optimum plant sample size was found to be 20-30 plants to estimate seed yield/plant, 30 plants for number of pods and seeds/plant, 10 plants for plant height and number of branches/plant. Biological and straw yield/plant were not accurately estimated from samples and they should be estimated from whole plots. Similar trend for sample size was obtained in both varieties in most studied characters, which exhibited significant location x sample size interaction. Therefore, further studies should be carried out in various locations rather than in a single location to remove the location effect