With the aims of accounting for the effects of the internal thermal resistance of the sample holder on the parameters of recorded DTA curves, and of estimating the difference between the instrument with a thermally insulated sample holder and the gradientless model, a novel two-point method of differential thermal analysis has been developed. Its essence is that two thermoanalytical curves are recorded simultaneously, with the differential thermocouple at central and side positions relative to the sample. The theory of the method has been elaborated, and formulae are derived which allow quantitative estimation of the thermal resistance of the sample holder, depending on the manner of packing and on the state of the sample in the holder, and which also indicate the optimum manner of packing. If the packing is not dense and not uniform, the thermal resistance of the holder increases and the accuracy of instrument calibration at the tail-end of the differential curve decreases by 10–20%. Through introduction of a correction term into the formula, this effect can be eliminated. A basic formula is given for DTA calculation in the general case of a sample holder with non-zero internal thermal résistance.