A cold-sensitive mutation in the structural gene for a minor phage T4 capsid protein (p20) leads to formation of heads containing p20 and cleaved head proteins and empty of DNA. Such heads can be filled with DNA and converted to active phages in vivo upon shift to high temperature. It appears that p20 has two distinct roles in head assembly: first, in construction of the prehead shell (blocked by ts and am mutation) and, second, in DNA packaging (blocked by cs mutation). The latter function is closely associated with gene 17 product, previously known to be required for DNA packaging. Temperature shift studies of cs-s double mutants and other observations allow determination of phage functions required for DNA packaging. Contrary to previous proposals, we find that T4 DNA packaging is not directly coupled to and can follow DNA synthesis, protein cleavage, prehead core removal, and gene 21-mediated cleavage-induced increase in head volume. Our evidence suggests that an altered head assembly pathway exists and that DNA packaging is probably initiated by DNA-capsid (p20) (p23) (5). In this paper we present evidence suggesting that neither of these pathways is correct. The evidence is also against the proposal that filling of T4 heads with DNA depends upon concurrent DNA synthesis (6).These questions are important because they relate to the function of protein cleavage in virus assembly and the unknown mechanism of DNA condensation. It was initially thought that basic DNA-binding internal proteins of the mature head might condense the DNA, but these proteins proved to be nonessential (1, 7). It was then proposed that DNA packaging was coupled to cleavage of core proteins, so that the generation of acidic internal peptide core remnants might cause repulsive collapse to {-like DNA (8, 9). Similarly, packaging could be coupled to exit of core proteins from the prehead in phages lacking protein processing (10). Another general idea is that packaging is coupled to the expansion of the capsid shell which accompanies protein processing and head filling (11, 12).In this paper we characterize a cold-sensitive (cs Fig. 4k) and was not associated with polyhead production, but gave rise to empty head structures (Fig. 1). The phenotype of this cs2O mutation is similar to that of gene 17 mutations. Indeed, the cs20 mutation and gene 17 function appear very closely interrelated because of (i) similar morphological phenotypes (empty head and ghost production); (ii) accumulation of similar sized concatameric T4 DNA (unpublished); (iii) Thin sections of bacteria infected with cs20 at nonpermissive temperature reveal many empty heads in the interior of the cell (Fig. 1). These clearly differ from the protein core-containing membrane-associated Tau-particles which are also seen in small numbers and are known to be intermediates in head formation (2, 19). The cs20 heads appear to be empty of DNA or core material, and to be of roughly the same size as the occasional head fully packaged with DNA seen in growth under slightly le...