| Abstract | We observed NGC 3379 and nine Virgo elliptical galaxies at 2.16 microns using a NICMOS3 256 X 256 detector array. We have measured the surface brightness fluctuations in the K-band for all but one Virgo galaxy (an E/S0), thus extending the previous measurements in K by a factor of 3. We find for NGC 3379 that the apparent fluctuations magnitude m^bar^_K_ = 24.48 +/- 0.11, and we find no evidence at the 0.20 mag level for radial gradients in the apparent fluctuations magnitude. In the Virgo Cluster, we find that NGC 4365 has a similar fluctuation magnitude to the other galaxies in the sample, which is in contrast to previous I-band fluctuations observations which put it in the background W cloud; we note that this discrepancy could also be the result of a different stellar population. We also find an anomalous bright apparent fluctuations magnitude for the dwarf, low-surface brightness elliptical NGC 4489, which might be attributable to its lower signal-to-noise ratio (but is significant at the 3 σ level) or to an unusual stellar population. The average fluctuations magnitude for the Virgo galaxies (excluding NGC 4489) is m^bar^_K_ = 25.16 +/- 0.18, with a 1 σ_rms_ dispersion of 0.25 mag. When we assume the distance moduli to these galaxies from I-band surface brightness fluctuations measurements, we find that for the sample of Virgo galaxies (excluding NGC 4365 and 4489), plus NGC 3379 M31, and M32 (the latter two from Luppino & Tonry 1993), the absolute fluctuations magnitude M_bar__K_ = 5.74 +/- 0.18. We find an observed 1 arms dispersion in the measurements of M^bar^_K_ of 0.20 magnitude, which is nearly that of our typical uncertainty of 0.18 mag in each measurement of M^bar^_K_. The theoretical predictions of Worthey (1993) for M^bar^_K_ as a function of both (V - I) color and Mg_2_ index are shown to be consistent with the observations, differing only by +0.31 and -0.09 mag, respectively, from our mean measurement of M^bar^_K_ = - 5.74 +/- 0.18. This difference is well within the quoted uncertainties of the models (0.4 to 0.5 mag) and our measurements. The simple stellar population models of Buzzoni (1993) are shown to be too faint in M^bar^_K_ by ~ 0.5 mag, which we suggest to be due to their exclusion of an empirically based M-giant population. We suggest that the origin of the dispersion in the measurement of M^bar^_K_ may be due to variations in metallicity; larger samples of early-type galaxies should be able to determine if there is such a relationship between metallicity and M^bar^_K_. We present models of M^bar^_K_ based upon the addition to the stellar population of an anomalous asymptotic giant branch population contributing 10% of the light at 2 microns and show that this produces an absolute fluctuations magnitude 0.7 to 1.5 mag brighter. Such a population may be consistent with the observations of M^bar^_K_ in M32 by Luppino & Tonry (1993), but could only be present in NGC 4365 and 4489 of our sample, the two galaxies which are measured to have brighter m^bar^_K_ relative to m^bar^_I_. Finally, we suggest that with the typically better seeing at the K band (compared with the optical), and with the brightness of the fluctuations at this wavelength, measurements of distances to greater redshifts are possible than with optical measurements of surface brightness fluctuations, although the problem of calibrating the relationship between M^bar^_K_ and (V - I) color or Mg_2_ index needs to be solved in order to reduce the uncertainties of such distance measurements. |
