The existence of surface sialytransferases that use cytidine monophosphate (CMP)-sialic acid as substrate has been postulated in previous studies. This is based on the assumption that if whole, viable cells can catalyze the transfer of sialic acid from CMP-sialic acid to endogenous acceptors, then the transferases carrying out the reaction must be on the cell surface, provided that (1) CMP-sialic acid does not enter the cells, and (2) CMP-sialic acid does not break down outside the cells, yielding free sialic acid which then may enter the cells, in amounts large enough to explain the incorporation. We now report evidence showing that after incubation of intact NIL, BHK, and 3T3 fibroblasts with CMP-sialic acid, at least 78% of the sialic acid incorporated by these cells is the result of free sialic acid uptake. When cells growing in a monolayer were incubated with a mixture of CMP-[14C]sialic acid and [3H]CMP-sialic acid with a ratio of 3H/14C=0.60, this ratio was found to be markedly increased in whole cells. Chemical analyses of the radioactive species in the incubation medium showed that a considerable portion of the radiolabeled sugar nucleotide had broken down to cytidine, phosphoric acid, and sialic acid. Upon incubation of cells with doubly labeled sugar nucleotide in the presence of a large excess of both nonradiolabeled cytidine and sialic acid, the cells incorporated less than 6% of both isotopes. Incubation of cells with a mixture of CMP-[14C]sialic acid and [3H]sialic acid resulted in only 20-40% of the radioactivity within the cells being membrane bound, and 70-90% of this incorporation could be inhibited by the addition of 10 mM azide to the incubation medium. The possibility that a small fraction of the total incorporation of sialic acid by these cells is due to surface sialytransferases cannot be completely ruled out. The uptake of free sialic acid by these fibroblasts is concentration dependent and a portion of it is incorporated into glycoproteins and glycolipids. Considerable loss of cell integrity was observed when fibroblasts grown on plates were removed by (ethylenedinitrilo)-tetraacetic acid or trypsin and subsequently incubated in buffer, indicating that these preparations are not suitable for intact cell studies.