Caveolae are vesicular organelles with a characteristic uniform diameter in the range of 50-100 nm. Although recombinant expression of caveolin-1 is sufficient to drive caveolae formation, it remains unknown what controls the uniform diameter of these organelles. One hypothesis is that specific caveolin-caveolin interactions regulate the size of caveolae, as caveolin-1 undergoes two stages of self-oligomerization. To test this hypothesis directly, we have created two caveolin-1 deletion mutants that lack regions of caveolin-1 that are involved in directing the self-assembly of caveolin-1 oligomers. More specifically, Cav-1 delta61-100 lacks a region of the N-terminal domain that directs the formation of high molecular mass caveolin-1 homo-oligomers, while Cav-1 deltaC lacks a complete C-terminal domain that is required to allow caveolin homo-oligomers to interact with each other, forming a caveolin network. It is important to note that these two mutants retain an intact transmembrane domain. Our current results show that although Cav-1 delta61-100 and Cav-1 deltaC are competent to drive vesicle formation, these vesicles vary widely in their size and shape with diameters up to 500-1000 nm. In addition, caveolin-induced vesicle formation appears to be isoform-specific. Recombinant expression of caveolin-2 under the same conditions failed to drive the formation of vesicles, while caveolin-3 expression yielded caveolae-sized vesicles. These results are consistent with the previous observation that in transformed NIH 3T3 cells that lack caveolin-1 expression, but continue to express caveolin-2, no morphologically distinguishable caveolae are observed. In addition, as caveolin-2 alone exists mainly as a monomer or homo-dimer, while caveolins 1 and 3 exist as high molecular mass homo-oligomers, our results are consistent with the idea that the formation of high molecular mass oligomers of caveolin are required to regulate the formation of uniform caveolae-sized vesicles. In direct support of this notion, regulated induction of caveolin-1 expression in transformed NIH 3T3 cells was sufficient to recruit caveolin-2 to caveolae membranes. The ability of caveolin-1 to recruit caveolin-2 most likely occurs through a direct interaction between caveolins 1 and 2, as caveolins 1 and 2 are normally co-expressed and interact with each other to form high molecular mass hetero-oligomers containing both caveolins 1 and 2.