Presently, the question about the length of microtubules in the interphase cell became actual, since the parameters of dynamic instability of the plus end measured in vivo do not allow one to explain the rapid turnover of the long microtubule system. The problem may be solved if one of the following suppositions is assumed: either microtubules undergo rapid depolymerization from the minus end or they are on the average much shorter than it is usually considered. To check the last hypothesis, we have reconstructed microtubules using stereophotography of electron microscopic sections. Microtubules around the cell center in cultures of epithelial cells (kidney of pig embryo (PK) and bovine trachea (FBT)) and fibroblasts (MEF, primary mouse embryo fibroblasts, and L cells), as well as at the periphery of PK cells were studied. All in all, no less than 200 microtubules were found near the centrosome in each cell culture. From 2.5 to 8% microtubules were beyond the studied volume (4.0 x 5.5 x 1.5 microm). Most of microtubules in all studied cell lines were up to 1 microm and about 1/3 of them were 0.2-0.4 microm long. The mean length of microtubules surrounding the centrosome in different cell lines differed insignificantly and equalled 0.4-0.8 microm. In this case, the microtubules attached to the centrosome were on the average slightly shorter than the free ones. Thus, almost all microtubules around the centrosome are short, and the majority of those attached to it do not reach the cell periphery. A similar reconstruction of a part of the PK cell cytoplasm (10 x 35 microm) has shown that at the periphery, the mean length of microtubules is about 1.6 microm and most of them are 0.5 to 1.5 microm long. Thus, our data confirm the recent hypothesis of Vorobjev et al. (I. A. Vorobjev, T. M. Svitkina, and G. G. Borisy, J. Cell Sci. 110:2635-2645 (1997)) that most of microtubules in the cells are not connected with the centrosomes.