The nucleic acids of many plant viruses encode proteins with one or more of the following properties: an Mr of approximately 30,000, localization in the cell wall of the infected plant and a demonstrated role in cell-to-cell transport of infection. A progressive alignment strategy, aligning first those sequences known to be similar, and then aligning the resulting groups of sequences, was used to examine further the relatedness of the amino acid sequences of putative transport proteins of caulimoviruses, of proteins similar to the putative transport protein of alfalfa mosaic virus (A1MV) and of those similar to the tobacco mosaic virus (TMV) 30K protein. The strategy first identified regions in which multiple dipeptides of one group were similar to those of another group. The regions of similarity were brought into alignment by the conservative introduction of gaps. The positions of the introduction of gaps were adjusted to optimize similarity. Statistical significances of the resulting alignments, determined both by comparison with shuffled amino acid sequences and with the sequence alignment off-set by 1 to 15 residues in each direction, suggest that the amino acid sequences of the three groups of viruses are distantly related. Nevertheless, significant relationships between members of the caulimoviral group of sequences and members of each of the A1MV-like and TMV-like groups were found. These relationships and the analysis of the number of insertions/deletions between present sequences and a hypothetical common ancestor suggest that the sequences of the caulimoviral proteins are less diverged from the ancestor than either the A1MV-like or TMV-like proteins. The alignment identified common regions of predicted secondary structure and regions of similar hydropathy, regions possibly crucial for proper functioning of the proteins.