The authors are to be congratulated for applying backpropagation neural network (BPN) to determine the compressive strength and strain of circular concrete columns. However the discussers would like to point out the following. The authors have used the architecture of 7-2-2, 7-3-2, and 7-4-2. The determination of number of neurons in the hidden layers is based on trial-and-error procedure and the authors have limited the hidden neurons to four. The discussers have used sequential learning neural network (SLNN) originally proposed by Zhang and Morris (1998) for the experimental data given by the authors. It uses a single hidden neuron with the Sigmoidal learning law and linear learning law for input and output layers. Out of 38 data sets given in the original Table 2, the odd-numbered data is used for training and the even-numbered data sets for testing. The network is trained for 150,000 epochs with a learning rate of 0.6 and a gamma value of 0.000001 and using an orthogonalization procedure. The reader may refer to the paper by Zhang and Morris (1998) and Rajasekaran and Amalraj (2002) for further details. The procedure uses two networks separately; one for peak stress and the other for strain at peak stress. The network consists of seven input neurons with a bias neuron; one hidden neuron and one output neuron (for peak stress or strain at peak stress as the case may be). The error rate versus log