The complete nucleotide sequences were determined for three transcripts each encoding a different variant surface glycoprotein (VSG) of Trypanosoma (Nannomonas) congolense. The nucleotide sequence was determined also for a transcript encoding a fourth VSG, but this was truncated. The data obtained confirm absence of the canonical polyadenylation signal, lack of conserved sequence elements in the 3' untranslated region, and heterogeneity in the spliced-leader acceptor site in the T. congolense VSG transcripts examined. A comparison of the amino acids deduced from the nucleotide sequences of the four VSGs and those of other VSGs published previously reveals a strong conservation of several structural domains, particularly cysteine residues located throughout most of the molecules. The majority of T. congolense VSGs analyzed in this study resemble most the N-terminal cysteine residue domain type B of T. brucei, characterized by a cysteine residue located toward the N-terminal end, a cluster of cysteine residues in the central region, and at least three cysteine residues between positions 250 and 300 of the molecules. One of the VSGs analyzed, ILNat3.3, did not fit into any of the classification schemes proposed for the VSGs so far studied, and thus may represent a different class of these surface molecules. Unlike VSGs of T. brucei, the T. congolense VSGs have no cysteine residues at the carboxy-terminal end. These data now make it possible to predict general primary structural features of T. congolense VSGs.