Click for text of press release
First is the color composite of the cluster field, with cluster members marked and enlarged, covering the whole WFPC2 field using filters F450W (blue), F606W (green), and F814W (red). The faint z=2.4 cluster objects don't exactly jump out to the eye. Below is a GIF file; the full-blown products are available, first as an unlabelled composite PostScript file . The labelled color-composite image, with postage-stamp enlargements of the cluster candidates, was Fig. 1 of the Nature paper and may be found as a compressed PostScript file. (since the uncompressed file is 12 Mb, a real exercise in patience from our site). We also have a PostScript preprint of the Nature paper, with the other figures: 2 3 4 5 .
The cluster of high-redshift objects was found by comparing broad-band blue images (F450W, B) with narrow-band (F410M) images. The wavelength of this latter filter, originally included in HST to study molecular features in planetary atmospheres, matches the greatly redshifted Lyman-alpha emission line of hydrogen, which we knew to be present in at least 3 galaxies in this field from ground-based work. The broad-band image is the top half, showing the many galaxies in this direction that are mostly foreground and which do not have emission at the cluster redshift. Notice how they almost vanish in the narrow band, leaving emission-line galaxies at this redshift standing out quite strongly. This image uses 0.05" subpixels, a result of 2x2 dithering in the exposure stacks that went into it (thanks, John Biretta and the scheduling folks!). The GIF file is straight with no overlays; a PostScript file is also available which has been bordered and annotated.
Next is a montage of cluster members, comparing images in B and Lyman alpha. B is on the left in each case with the matching area in Lyman alpha on the right. I have kept our working object numbers to facilitate discussion. Each panel is 2.5" square. I tried to keep a consistent intensity mapping, though there are a few where the surface brightness is low enough that I had to jack it up. The GIF version is a screen snap and thus loses some fidelity - see the PostScript version for an original product. (This was done using SGI Showcase, which only lets me save internal or PostScript formats...) Number 6 is the radio galaxy 53W002 on which the observations were originally targeted. Several of these show substructure (particularly double lumps) which fits nicely in the merging scenario. For objects this faint, any detection at all in the narrow Lyman-alpha filter implies significant emission at the desired wavelength.
Here is Sam Pascarelle's cartoon history of galaxy formation. The original PostScript file is 4 Mb and took almost 15 minutes to render on my SGI, so here are some trashy screensnap GIF files for quick perusal:
To show more clearly how small these objects are when compared to today's luminous galaxies, here I compare some of them to the nearby spiral M101. The M101 image is an ultraviolet (1500-Angstrom) frame taken with the UIT payload on the Astro-2 mission; because of the cluster's large redshift, this is the same range of emitted wavelength that we see in the middle of the optical band. The circular field around M101 is 40 kpc in diameter, about 40 half-light radii of typical high-redshift objects (I took a Hubble constant of 80 km/s Mpc and deceleration parameter q0=1/2 for the scaling). As usual, the GIF image is a screensnap, the real thing lives in a PostScript file .firstname.lastname@example.org