We present B, V: I CCD photometry of about 40000 stars in four regions of t
he Fornax dwarf spheroidal galaxy down to V similar to 23.5, the largest th
ree-color data set obtained for this galaxy until now. The resultant color-
magnitude diagrams, based on a wide color baseline, show a variety of featu
res tracing the history of star formation of this dwarf galaxy. One of the
most distinctive features in our diagrams is the conspicuous young main seq
uence, indicating recent star formation until approximately 2 x 10(8) yr ag
o. A plume of stars brighter than the red HE clump, with (B - I) similar to
0.5, trace the helium-burning phase of the young population. A comparison
of the color and extension of this feature with model isochrones suggests a
relatively metal-rich population ([Fe/H]similar to -0.7) with age 300-400
Myr. This represents an important constraint for understanding the chemical
enrichment history of Fornax. An extended upper AGE tail and a prominent r
ed HE clump sign the presence of the well-known dominant intermediate-age p
opulation with an age range 2-10 Gyr, for which we have estimated a mean ag
e 5.4 +/- 1.7. About 0.2 mag below the red clump, an extended HE is indicat
ive of an old population. We show that blue HE stars may be present in the
outer regions. Together with previous detection of RR Lyrae, this provides
evidence for a minority field population that is as old and metal-poor as t
hat in the Fornax globular clusters. We have identified the AGE bump, a clu
stering of stars that occurs at the beginning of helium shell-burning evolu
tion, at a luminosity M-V similar or equal to -0.4. This is an example of t
he short-lived evolutionary phases that can be revealed in stellar populati
ons using adequately large star data samples, whose measurements provide po
werful tests of theoretical models.
Based on precise detection of the tip of the RGB in a selected RGB sample,
we measure a corrected distance modulus (m - M)(0) = 20.70 +/- 0.12. An ind
ependent estimate of the distance to Fornax was also obtained from the mean
magnitude of old horizontal branch stars, yielding a distance modulus (m -
M)(0) = 20.76+/-0.04, in good agreement with the distance estimated from t
he red giant branch tip and previous results. The large baseline of the (B
- I) colors together with the size of the stellar sample allowed us to anal
yze in detail the color distribution of the red giant strars. We rind that
it can be approximately described as the superposition of two populations.
The dominant component, comprising similar to 70% of the red giant stars, c
onsists of relatively metal-enriched intermediate-age stars. Its mean metal
licity is [Fe/H]=-1.39 +/- 0.15, based on a compar- ison of the fiducial lo
cus of the bulk or the Fornax red giants with the homogeneous Galactic glob
ular cluster set of Da Costa & Armandroff ( 1990). Once the younger mean ag
e of Fornax is taken into account, our best estimate for the mean abundance
of the bulk of the galaxy is [Fe/H]approximate to -1.0+/-0.15. The dominan
t intermediate-age component has an intrinsic color dispersion sigma(0)(B -
I) = 0.00 +/- 0.01 mag, corresponding to a relatively low abundance disper
sion, sigma([Fe/H]) = 0.12 +/- 0.02 dex. Further, there is a distinct small
population of red giants on the blue side of the RGB. While these stars co
uld be either old or young red giants, we show that their spatial distribut
ion is consistent with the radial gradient of old horizontal branch stars,
and completely different from that of the younger population. This unambigu
ously qualifies s them as old and metal-poor This result clarifies the natu
re of the red giant branch of Fornax, suggesting that its exceptional color
width is due to the presence of two main populations yielding a large abun
dance range (-2.0 < [Fe/H] < -0.7). This evidence suggests a scenario in wh
ich the Fornax dSph started forming a stellar halo and its surrounding clus
ters together about 10-13 Gyr ago, Followed by a major star formation epoch
(probably with a discontinuous rate) after several Gyr.