Biotin labeling of red cells was studied using different approaches to see if biotinylation is a useful label for determination of erythrocyte survival. Mouse red cells were labeled with biotin, either in vivo by injection or in vitro. In vivo labeled red cells were followed up in some mice without transfusing the labeled erythrocytes. Furthermore, in vivo labeled as well as in vitro labeled red cells were transfused into syngeneic mice. The biotin label allows an easy discrimination between labeled and unlabeled red cells during FACS analysis, and it is relatively stable for at least 50 days. All the three different approaches give similar results. Mean red cell life spans of in vivo or in vitro labeled red cells either transfused or followed up in vivo were between 44 and 52 days (T50 mean 23.9 days) when red cell destruction was assumed to be only a result of senescence. Mean red cell life spans were between 8 and 18 days (T50 mean 9.5 days) when a random destruction independent of red cell age was suggested. All the survival slopes are neither simple linear functions of time nor logarithmic functions, but they show an overlay of linear function by a logarithmic function where the components of both are unknown.