During sea urchin gastrulation filopodia are sent out by secondary mesenchyme cells (SMCs) at the tip of the archenteron in continual cycles of extension, attachment, and retraction. Eventually the archenteron ceases its elongation and its tip localizes to the animal pole region of the embryo (Gustafson and Kinnander, 1956, Exp. Cell Res. 11, 36-57; Dan and Okazaki, 1956, Biol. Bull. 110, 29-42). We have investigated the mechanisms and specificity of this localization by analyzing filopodial behavior and by experimental manipulation of the interaction of the archenteron with the animal pole region. When the tip of the archenteron nears the animal pole, some filopodia make contact with a well-defined locus within this region. Filopodia that make contact with the locus remain attached 20-50 times longer than attachments observed at any other site along the blastocoel wall. The SMCs bearing the long-lived filopodia eventually change their phenotype by flattening and spreading onto this region. Several lines of experimental evidence indicate that contact with the animal pole locus, or "target" region, is crucial for the change in phenotype of the SMCs: (1) the phenotypic change can be induced precociously by bringing the animal pole region within reach of the tip of the archenteron early in gastrulation. Precocious contact with other regions of the blastocoel wall does not induce a similar change. (2) The phenotypic change can be delayed by placing the animal pole out of reach late in gastrulation, resulting in artificial prolongation of exploratory behavior by filopodia. (3) Ectopic combinations of animal pole ectoderm and archenterons in fused multiple embryos and chimaeras result in attachment of archenterons to the nearest available target, and (4) freely migrating SMCs are observed to migrate randomly within the blastocoel, then stop at the animal pole and undergo the change in phenotype. Filopodia rapidly attach to the animal pole when the shape of early gastrulae is altered such that the animal pole is less than 35 microns from the tip of the archenteron, even though such attachments only occur in normal embryos at the 2/3-3/4 gastrula stage. Since it has previously been shown that the archenteron elongates autonomously to 2/3 of its final length (Hardin, 1988, Development 103, 317-324), it appears that autonomous extension of the archenteron is required to place filopodia close enough to the animal pole to allow them to interact with it.(ABSTRACT TRUNCATED AT 400 WORDS)