Field experiments with three seed sizes and three seeding rates of each of three barley varieties were conducted in 1959 and 1960 at Edmonton and Lacombe, Alberta. Data were obtained on emergence, culm counts, grain yields, and 1000-kernel weights. Although there was no effect of seed size on emergence, the culm counts and yields were highest for large seed and lowest for small seed over all rates, varieties, and tests. The differences between yields from large and small seed decreased slightly as the seeding rate increased, but the seeding-rate × seed-size interaction was not statistically significant. There were varietal differences in yield response to seeding rates. It was, therefore, concluded that seeding rates in test plots should conform to those used in farming practice. The rate of seeding had a significant influence on 1000-kernel weights. The lowest rate produced the heaviest kernels and the highest rate produced the lightest kernels in each of four tests. Tillering capacity alone was a poor indicator of resulting yields. Results from further experiments, involving four seed categories of nine varieties in two tests, corroborated the above findings.