The procyclic form of Trypanosoma brucei exists in the midgut of the tsetse fly. The current model of its surface glycocalyx is an array of rod-like procyclin glycoproteins with glycosylphosphatidylinositol (GPI) anchors carrying sialylated poly-N-acetyllactosamine side chains interspersed with smaller sialylated poly-N-acetyllactosamine- containing free GPI glycolipids. Mutants for TbGPI12, deficient in the second step of GPI biosynthesis, were devoid of cell surface procyclins and poly-N-acetyllactosamine- containing free GPI glycolipids. This major disruption to their surface architecture severely impaired their ability to colonize tsetse fly midguts but, surprisingly, had no effect on their morphology and growth characteristics in vitro. Transmission electron microscopy showed that the mutants retained a cell surface glycocalyx. This structure, and the viability of the mutants in vitro, prompted us to look for non-GPI-anchored parasite molecules and/or the adsorption of serum components. Neither were apparent from cell surface biotinylation experiments but [H-3]glucosamine biosynthetic labeling revealed a group of previously unidentified high apparent molecular weight glycoconjugates that might contribute to the surface coat. While characterizing GlcNAc-PI that accumulates in the TbGPI12 mutant, we observed inositolphosphoceramides for the first time in this organism.