An important aspect of dating problems is the question of the number of samples which are necessary and sufficient and which should be dated for each event. This number may be frequently somewhat arbitrarily chosen, but should be more often based on scientific considerations and the need for accuracy. In some cases a single sample is enough, but in general more samples are needed. The main reason for dating more than one sample is often the possibility of contamination or systematic error. It may be of little use in obtaining high accuracy in sample counting if other errors are more predominant than the purely statistical ones. It is often difficult for the submitters to select an adequate number of samples at the beginning of a dating project. The most satisfactory result may be obtained when additional samples are added during the work. Close collaboration with the dating laboratory seems to be an advantage. A few examples, in particular the dating of a Norse settlement in Canada, illustrate some of the problems involved in obtaining optimal accuracy.