What determines the length of a cell — its type or some other factor? To tackle this fundamental question, Margarita Kharitonova and Jury Vasiliev have monitored growth of cells along narrow adhesive channels created by scratching glass coverslips coated with non-adhesive plastic. One might expect that cells would grow longer when forced to be narrower by such channels. This is true of epithelial cells — but not fibroblasts, in which a `length control' mechanism limits their growth. So what is its basis? Kharitonova and Vasiliev show that treating fibroblasts with the microtubule-disrupting drug Taxol converts them from elongated polarized cells containing straight actin bundles to discoid cells that have circular actin bundles; significantly this also abolishes length control (see p. 1955). By contrast, when they convert the discoid epithelial cells into elongated cells containing straightened actin bundles by adding scatter factor, the cells become subject to length control. Length control thus seems to be a feature of a cell's cytoskeletal organization and shape rather than of particular cell types. Kharitonova and Vasiliev propose that it is cooperation between microtubules and actin stress fibres running parallel to the axis of the cell that underpins this size control mechanism.