326 Highlights in modern observational cosmology
compute the K-corrections). Such assumptions reflect observations of the nearby
universe but are still affected by some uncertainty, therefore it is not uncommon
to find in the literature NE models which differ by∼50%. This uncertainty will
be drastically reduced when the 2dF and SDSS surveys are completed.
A clear trend is apparent in figure 11.3. At blue wavelengths the observed
counts exceeds the NE predictions by as much as a factor three, a problem
which was recognized in the first deep surveys and which has become known
as thefaint blue galaxy excess. Such an excess progressively disappear at longer
wavelengths. Observations in blue filters are sensitive to late type, star-forming
galaxies with young stellar populations. Therefore, it had already become evident
in the early 1990s (e.g. Elliset al1996) that this is the galaxy population which
has undergone most of the evolution (in luminosity and/or number density) out
toz ∼1, i.e. the last 50% of the life of the universe. The first deep redshift
surveys (Lillyet al1995) confirmed this scenario directly measuring a significant
evolution of the LF for the ‘blue population’ out toz 0 .7, while revealing no
significnt evolution for the ‘red population’ consisting of galaxy types earlier than
an Sbc (see figure 11.4). Red wavelength observations, particularly in the K-band
(λ 0 = 2 μm), collect rest-frame optical light out toz∼3, thus probing old,
long-lasting stellar populations in distant galaxies (i.e. earlier types). All these
observations (see also Cowieet al1996) have shown a remarkable increase in the
space and/or luminosty density of star-forming galaxies with redshift. However,
interpreting these results, and understanding the physical processes responsible
for this evolutionary pattern, has remained a difficult task.
In this respect, HST observations have driven us a big step forward by
allowingintrinsic sizesandmorphologiesof distant galaxies to be measured.
The combination of angular resolution (0. 05 ′′) and depth has also pushed these
studies well beyondz =1. As an example, in figure 11.5 we show number
counts for different morphological types as directly determined by the HDF-
Nimages(Driveret al1998). Along with NE model predictions (full lines),
passive evolution modelsare also shown. The latter are constructed using spectral
synthesis models (e.g. Bruzual and Charlot 1993), assuming a formation redshift
(generally varying by type), and a star formation history (with a given initial mass
function, IMF). As an example, in figure 11.6 we show the evolution of the SED
of a 3 Gyr burst stellar population over approximately a Hubble time. This model
well reproduces the evolution of an early type galaxy. The UV luminosity declines
rapidly after the end of the burst of star formation, as hot O and B stars burn off
the main sequence and the population is more and more dominated by red giants.
In general, passive evolution models are characterized by luminosity
evolution, which is the result of letting the stellar populations evolve with a
pre-defined star formation history, without including any merging. Figure 11.5
confirms that morphologically selected early types show little (simple passive)
evolution to faint magnitudes, and hence to relatively high redshifts. Counts of
intermediate types (i.e. spiral-like galaxies) are broadly consistent with passive,
luminosity evolution models, whereas later types and irregulars are not fitted by