Notes. Identifier (ID) of our candidates (Col. 1), spectral indices presented by Gizis (1997) (Cols. 2–5), classification based on the scheme from Lépine et al. (2007) (Col. 6), spectral type adopted using SDSS spectral templates (Col. 7), telescope where the spectrum was obtained (Col. 8), spectroscopic distances with their errors (Col. 9), heliocentric velocities (Col. 10) compared to spectral templates in our sample (ID = 29, 11, and 7 as sdM, esdM, and usdM templates, respectively), except for the three L-type subdwarf where we used the sdL3.5 subdwarf (Sivarani et al. 2009; Burgasser et al. 2009) and space motions for sources with revised proper motions in both directions. Some candidates appear twice because they have spectra from the SDSS database and our own spectroscopic follow-up. (1) Uncertainties on the distances take into account the error on the J-band magnitude of our target and the error on the trigonometric distances of the subdwarf templates listed in Table 1. We computed the minimum and maximum distances and quote the largest error. (2) For ID =19 we used the J-band absolute magnitude (M_J) of a M4.5 and sdM5.0, yielding distances of 162.4 ± 20.3 pc and 308.9 ± 44.3 pc, respectively. (3) For ID = 68, we list the distances assuming a spectral type of sdM6.0. If we consider the esdM5.5 classification, we find a distance of 322.0 ± 36.0 pc. For IDs = 98 and 100, we list the distances for the metal-poor case. If we assume that both objects are solar-metallicity M6.5 dwarfs, we find spectroscopic distances of 340.9 ± 24.0 pc and 144.4 ± 5.4 pc, respectively. (4) For objects whose spectral types are quoted as intervals, we used the earliest spectral types implying upper limits on the distances. extremes for the distance estimates without including the uncertainty of half a subtype.