This report is the second part of serial experiments which are being made to survey the effects of welding procedures on the shrinkage, distortion and residual stresses of welded joint. The effects of the combination of the welding direction, chipping and flame gouging, combination of the electrodes of various diameter, type of electrode, root gap and size of bevel on the transverse shrinkage of constrained butt joints are studied in this thesis. The ring form constrained specimens were used. Six serieses of experiment were performed The experimental conditions for each series are listed in Table 2-1. The measurement of shrinkage were performed after the welding of each layer. The main results obtained are summarized as follows: 1) The Value of Mean Shrinkage In the multiple layer welding, there exists a linear relation between the shrinkage and the logarithm of the rod consumption per the unit weld length. The effect of welding procedures on the mean shrinkage were investigated by this theorem. The main results obtained are as follows:-- The root gap and the size of bevel have remarkable effects on the shrinkage among other procedures. The value of shrinkage does not change by the chipping, and increases by the re-welding. For this reason the shrinkage at the final stage increases by the chipping and re-welding. About the effect of chipping and re-welding we obtained a simple formula to estimate it. On the other hand, the shrinkage increases by the flame gouging because of the supply of heat during flame gouging. 2) The Form of Shrinkage The shrinkage due to welding is not uniform along the weld line. As the form of shrinkage is nearly linear, we investigated the shrinkage as the combination of the uniform shrinkage and the rotation, About the effect of welding procedures on the rotational component of the shrinkage we obtained the following results:-- The form of shrinkage at the welding of the first layer greatly depends uppon the welding direction. In the multiple layer welding, the value of the rotational component gradually changes approaching to a certain value which originates the minimum strain energy. The existence of this fact has been already estimated by the preceding experiments related in Part I. We ensured it by the further investigation.