NEW GLOBULAR-CLUSTER CANDIDATES IN THE INNER REGIONS OF M-31 AND THE PROJECTED DENSITY PROFILE OF THE CLUSTER SYSTEM

A new search for globular clusters carried out in the inner regions of M 31 (within about 1 kpc from the nucleus, approximately 5.5 arcmin) has yielded 3 Bo-class A, 1 B, 20 C, and 20 D new candidates never inc luded in any previous list (see Battistini et al. 1987), and has allow ed us to revise the classification of all those already detected in th e same area. The total sample of possible candidates ever proposed so far has then been carefully revised, leading to the definition of the ''current best sample'' of M 31 cluster candidates including 341 objec ts. By considering the numbers of globular clusters detected in M 31 a nd in the Galaxy over annuli at increasing galactocentric distances, a total population ratio of about 2.5 +/- 1.0 (conservative error) has been adopted, and the total number of M 31 clusters outside a radius o f approximately 20 kpc should be approximately 65. In particular, 15-2 0 new clusters should be detected in the annulus 20-30 kpc, partially scanned by the available surveys, but the task is difficult since most of them should be fainter than V = 18 and Palomar-like. A second samp le, the ''adopted best sample'', thought to be sufficiently uncontamin ated, homegeneous, and complete (up to about 25 kpc from the M 31 nucl eus and down to V approximately 18) and including 298 clusters, has th en been defined to study the projected cluster density profile and to address the long-standing question of the so-called missing globulars with respect to the centrally extrapolated R1/4-law. The analysis of t he spatial distribution has shown that: i) Adopting a R1/4-law, the fi tted effective radius of the system, R(ef), is highly sensitive to any variation of the adopted sample but it does not differ significantly from the observed value, R(eo) approximately 4.59 kpc if a sufficientl y unbiased sample is used. However, various tests show that the propos ed R1/4-law is not necessarily the best fitting description of the obs ervations. ii) Assuming R(e) = R(eo), a best fitting procedure based o n R1/m-laws yields m approximately 1.6 as the optimal solution. This w ould imply that the spatial distribution of the globular clusters in t he central regions of M 31 is much flatter than the behaviour typical of the R1/4-law. iii) Fitting the data with R(-n) laws turns out to be worse for any value of n, the best being n = 0.9 if equal weights are assigned to all radial bins. iv) Various tests show that the best fit ting description of the observations, the quality of the fit, and the deduced parameters are significantly depending on the adopted errors a nd weights. v) If the validity of the R1/4-law is imposed outside a fi xed galactocentric distance (3,2,1,kpc), a residual central flattening persists even after this new search. vi) The same procedure applied t o the sample of Galactic globular clusters leads to similar results su ggesting, if confirmed, that the effective radius of the visible halo formed by the stellar spheroidal component of the parent galaxy is sma ller than that of the globular cluster system. However, the separation of true halo globulars from other globular clusters in M 31 is proble matic and this may still affect the results.