The paper makes an attempt to test empirically the validity of the marginal productivity hypothesis of wage determination in the case of the Indian economy by using the data on the organised manufacturing sector. The broad methodology followed in the study consists in estimating the Cobb-Douglas production function for Indian industries on the basis of the time series (1946-64) data and also the cross-section data for years 1960 & 1964 and in turn deriving the series of estimated value of marginal product of labour from the estimated production function. By regressing the observed wage rate on the estimated marginal product of labour, the linear relationship between the two is then estimated and tested. A few other test criteria such as the Douglas criterion are also applied. The main finding of the study is that wages paid in Indian manufacturing industries do not reflect the corresponding marginal productivity of labour. Tracing the divergence between the two to the market imperfections, an attempt is also made on an experimental basis to estimate the implicit elasticities of supply of labour in a few selected industries.