The objective of this work was to determine the minimum number of replications and seeds per replication to perform the germination and seed vigor tests with coriander seeds. Two seed lots were compared in terms of water content, vigor and viability. Eight hundred seeds per lot were used. Values of germination first count, germination, germination speed index and mean germination time were analyzed. Sample size scenarios were developed using different combinations of number of replications (from 2 to 10) and the number of seeds per replication (from 20 to 80) by means of a resampling with replacement technique. The reference sample consisted of four replications of 50 seeds, as it is commonly used in researches with this species. Determination of the minimum number of replications and seeds was based on comparing the 95% bootstrapped confidence intervals for the index of variation (CV/n0.5) of each scenario with the confidence interval of the reference sample. It is reasonable to reduce the number of seeds per replication from 50 to 30 in order determine germination first count, germination and germination speed index. Forty seeds per replication are recommended to quantify the mean germination time. However, reductions in the number of replications can affect negatively the accuracy of germination and vigor tests.