We describe the identification and characterization of two genes and their gene products responsible for determining receptor binding and pilus length in type 1-piliated Escherichia coli. One gene, pilE, conferred the ability of piliated cells to agglutinate guinea pig erythrocytes. The other gene, pilF, determined pilus length, in that mutants having lesions in pilF had very long pili. The two genes were detected after TnS mutagenesis of a cloned segment of DNA that normally complemented a pilE lesion in the chromosome. Thus, lesions in pilE or pilF on the cloned segment resulted in mutants having the PilE-phenotype (piliated but unable to agglutinate erythrocytes). Introduction of the plasmid-encoded mutant alleles of pilE and pilF into the chromosome followed by electron microscopic examination of the mutants showed that only lesions in pilF conferred the striking increase in pilus length. Mutations in pilF could be complemented in trans by the original cloned segment to produce cells with parental-length pili. Minicell transcription and translation of the cloned pilE and pilF genes having representative TnS insertion mutations showed that the pilE gene product was a protein of ca. 31 kilodaltons and that the pilF gene product was a protein of ca. 18 kilodaltons. We believe that the pilF gene product may act as a competititve inhibitor of pilus polymerization. Thus, pilus length may be controlled by the ratio of pilin to pilF gene product present within the cell.Type 1 pili of Escherichia coli provide a convenient model for studying the control and assembly of supramolecular extracellular structures. Also, they provide a model for cell-cell interaction, since type 1 pili are required to mediate attachment of E. coli to a variety of eucaryotic cells through a mannose-sensitive interaction with a receptor on the eucaryotic cell (4).One interesting facet of the assembly of type 1 pili concerns how the cell monitors the number of pili it has and how long those pili are. These two considerations would seem to be important for the cell to conserve energy. However, since pili are on the outside of the cell, it is unclear how the cell would keep track of these two parameters. It is apparent from recent work that at least one mechanism exists for controlling the number of pili per cell. This fact is indicated by the existence of hyperpiliated mutants (13). Since hyperpiliated mutants also have longer pili (13), one might infer that the cell has a way to monitor pilus length also. However, no such pilus length mutants have been isolated.Another interesting feature of type 1 pili concerns their involvement in receptor-ligand interactions. Recently, it has become evident that while type 1 pili are required for receptor binding, their presence is not sufficient for binding (8). This fact is indicated by the existence of fully piliated mutants that fail to agglutinate guinea pig erythrocytes (8).