The use of cell substrates to propagate viruses or recombinant plasmids for vaccine productions has been the subject of long evolutionary conflict, primarily from the standpoint of product safety, and especially from the viewpoint of cancer induction. The present concern is for safety of vaccines made using transformed or neoplastic mammalian cells that may contain endogenous contaminating viruses or integrated gene sequences from oncogenic viruses. There is also concern for use of plasmid vectors employing promoter elements from oncogenic viruses. The principal concern for safety lies with retention of residual DNA in the vaccine, especially since induction of cancer is a single-cell phenomenon, and a single functional unit of foreign DNA integrated into the host cell genome might serve to induce cell transformation as a single event or part of a series of multifactorial events. Current proposed standards for vaccines would permit contamination with up to 100 pg of heterologous DNA per dose. This is equivalent to about 10(8) "functional lengths" of DNA. Total safety would seem to require complete absence of DNA from the product. While preparation of biologicals used to treat serious disease might demand the use of mammalian cells, this is not the situation with vaccines that are given prophylactically to persons who might be given equally efficacious vaccines produced in bacterial cells or in yeast that have attributes for greater safety. Careful assessment of safety and risk vs. benefit of continuous mammalian cell-produced vaccines should be made by technically expert scientists in the relevant disciplines and a consensus needs to be evolved in the scientific community at large.