The recently discovered galaxy population at z ~ 3 [40] and the UV-dropouts in the Hubble Deep Field [38] have re-opened the debate on the SF and metal production histories of the high-redshift Universe (see Madau et al. 1996). One of the current issues is to understand which fraction of the total high-redshift SF these galaxies represent (e.g., Pettini, this Conference, [13, 25]). Here, the potential impact of dust reddening on the estimates of the high-redshift SF and metal production rates are briefly discussed (see, also, Meurer et al. 1997, and, for a detailed analysis, Dickinson 1998).
By the nature of the detection criterion
[41],
all the z 
3 galaxies have intense
UV emission, tracer of recent massive SF
[40,
16].
From the spatial extension of the UV emission (a few kpc,
[16]),
the galaxies have been forming massive stars
for at least 100 Myr and probably more; this is the amount of time required by
the 'SF wave' to cross a typical galaxy scale, 1 kpc, if it travels at
the sound
speed, v 
10 km/s. The
spectrum of a dust-free galaxy which is forming
stars at a constant rate since Td = 100 Myr has
~ -2.45,
and, if Td = 1 Gyr,
~ -2, with little
dependence on the metallicity down to 1/10 solar
[22,
5,
7]).
However, the UV indices measured in the high-z galaxies are redder than these
values; the observed average UV index of the distant galaxies is
-1.1
([28]).
Taking into account corrections for the Lyman Forest absorption, the
galaxies UV spectral energy distributions become slightly bluer,
-1.2 to
-1.5 (cf. Dickinson 1998),
but still too red to be compatible with on-going
constant-rate (or slowly decreasing) SF. To reconcile the observed
average UV
index with the expectation for a star-forming galaxy, three possible scenarios
(or a combination thereof) can be invoked: 1) the massive end of the Initial
Mass Function of the z = 3 galaxies is steeper than the present-day IMF; 2) the
z = 3 galaxies are 'aged star-forming galaxies': the SF happened in the
past and
massive stars are no longer being formed in large quantities; aging reddens
the UV index of a stellar population; 3) the UV emission of the z = 3
galaxies is
reddened by dust. Aging and reddening are particularly important, because
both scenarios suggest that the SF and metal production rates inferred from
the observed UV emission of the high-z galaxies are underestimates of the
global SF and metal production rates.
is independent of the
environment
[27,
39].
Invoking a different IMF between the high redshift and
the present-day galaxies thus requires an explanation of why the IMF has
changed with time. In this regard, it should be remembered that the
metallicity of high-z Damped Ly-
Systems is not drastically lower than the values
observed today in Blue Compact Dwarf galaxies. The shape of the IMF is
thus unlikely to have changed from z ~ 3 to the present, though the issue is
still open.
150 Myr
must lapse before
-1.2. At this stage, the UV
flux around 1500 Å is about f = 55 times lower than the value at t
100 Myr.
A Td = 1 Gyr star-forming population would need Ta
20-50 Myr to reach
-1.2. In this case, the
decrease in UV flux relative to the peak
is about a
factor f ~ 4-6. If the observed galaxies are located at z = 3 (3.5), the
peak of SF
(t Td) would occur at z
> 3.12 (3.7) in an open Universe
(H0 = 50 Mpc/km/s
and q0 = 0.0) and at z > 3.25 (3.9) in a flat Universe
(q0 = 0.5). This peak has not been observed (e.g.,
Madau et al. 1996),
although the incompleteness of
the samples may play a large role here; of course, the redshift of the peak can
be pushed to higher values if the SF doesn't interrupt abruptly after
Td (as
in our simplified model), but is a decreasing function of time.
0.15 is
needed to change the UV index from -2 to -1.2. This
modest color excess implies a UV attenuation A(1600 Å)
1.65. The intrinsic
UV emission, and corresponding SFR, is then underestimated on average by
a factor f ~ 5. Any stellar population younger than 1 Gyr will have bluer
intrinsic UV spectra, and will imply larger reddening corrections (cf.
Meurer et al. 1997).
A simple argument can be used to infer that the amount of metals produced in the z = 3 galaxies by the observed SF event are comparable for the aging and the reddening scenarios. The injection of metals into the ISM is proportional to the total number of massive stars produced. If the IMF is the same in the two scenarios, the number of massive stars generated by the SF event is proportional to the observed SFR, sfrobs, to the factor f by which sfrobs underestimates the peak SFR, and to the duration of the burst, Td. The quantity sfrobs comes from the observed UV emission and is the same in both scenarios. In the aging case, f x Td = 55 x 100 Myr and 4 x 1 Gyr, respectively; in the reddening case, f x Td = 5 x 1 Gyr. Therefore, both scenarios predict the same amount of metals produced at z ~ 3-4 over the lifetime of the SF event. More sophisticated models for the stellar populations and evolution of the high-z galaxies would probably still give numbers in this ballpark. Metal production thus does not appear to be a good discriminant between aging and reddening in the high-z galaxies.
Both reddening and aging predict that the observed UV emission from the distant galaxies underestimates the peak SFR; there is an important difference between the two, though: in the first case, the underestimated quantity is the 'current SFR' at z ~ 3: in the second case, the quantity is the 'recent past SFR' a redshift beyond 3. Probably the most direct way to prove whether the high-z average UV spectra are red because of dust obscuration or aging will be to measure the intensity of the hydrogen Balmer lines: these are more directly related to the number of ionizing photons than the UV emission, and can help constrain the average age of the massive star population.