This paper examines the time series properties of the domestic agricultural terms of trade, the external net barter terms of trade and the ratio of the agricultural to the external terms of trade in Bangladesh with annual data for the period 1952-2005. The ADF and the KPSS tests results suggest that the time series of these variables do not have a unit root. The Perron (1989) test results, which makes allowance for a structural break in the series during 1972-1975 because economic and political shocks, suggest that although the domestic agricultural terms of trade does not have a unit root, the null hypothesis of a unit root in both the external terms of trade and the ratio of the agricultural to the external terms of trade cannot be rejected at the conventional 5 per cent level. Having obtained such mixed results, these series are assumed trend-stationary for examining any dynamic relationship between them Accordingly, the domestic agricultural terms of trade and the external terms of trade are detrended and a Granger-causality test is conducted between them. Although the test results are sensitive to the lag-length, they suggest that there is a bi-directional causality between them. This finding casts doubt about the view that the domestic agricultural terms of trade in Bangladesh can be considered a policy instrument that remains under the control of policy makers, who might have used it to 'squeeze agriculture' for industrialisation under an import-substituting strptegy of development until the early 1980s.