We have analysed, for the first time, the clustering properties of Wolf-Rayet (W-R) galaxies, using a large sample of 846 W-R galaxies selected from the Data Release 4 (DR4) of the SDSS. We compute the cross-correlation function between W-R galaxies
and a reference sample of galaxies drawn from the DR4. We compare the function to the results for control samples of non-W-R star-forming galaxies that are matched closely in redshift, luminosity, concentration, 4000-AA break strength and specific star formation rate (SSFR). On scales larger than a few Mpc, W-R galaxies have almost the same clustering amplitude as the control samples, indicating that W-R galaxies and non-W-R control galaxies populate dark matter haloes of similar masses. On scales between 0.1--1$h^{-1}$ Mpc, W-R galaxies are less clustered than the control samples, and the size of the difference depends on the SSFR. Based on both observational and theoretical considerations, we speculate that this negative bias can be interpreted by W-R galaxies residing preferentially at the centers of their dark matter haloes. We examine the distribution of W-R galaxies more closely using the SDSS galaxy group catalogue of Yang et al., and find that $sim$82% of our W-R galaxies are the central galaxies of groups, compared to $sim$74% for the corresponding control galaxies. We find that W-R galaxies are hosted, on average, by dark matter haloes of masses of$10^{12.3}M_odot$, compared to $10^{12.1}M_odot$ for centrally-located W-R galaxies and $10^{12.7}M_odot$ for satellite ones. We would like to point out that this finding, which provides a direct observational support to our conjecture, is really very crude due to the small number of W-R galaxies and the incompleteness of the group catalogue, and needs more work in future with larger samples.
