The extreme gene, also called extreme non-agouti, is represented as ae and is the lowest recessive allele on the A-locus. It's called "extreme black" or "super black" from time to time. So if a normal black is a/a, an extreme black is ae/ae. You have to have the gene to introduce it and it is unrelated to the extension locus (which uses "e" as its initial and causes confusion with ae sometimes). Since everything on the A-locus (Ay, Avy, Ahvy, Aw, A, at, am, a) is dominant to it, you may have it lurking around in any population of mice and not necessarily know it. Because it causes an overabundance of pigment, mice who display two copies of this gene are slightly smaller than other mice, as a general rule. Just ask any breeder of show blacks--their mice are usually the smallest at any show!
Most good blacks (whether a/a U/* or ae/ae) carry
nothing else. This is a very important point to remember. You would never breed a good black to a chocolate or blue or spotted mouse, etc. Black is the only acceptable outcross, period.
Umbrous, represented U, is a gene which only shows some of the time and only on some varieties. It's generally classified as a dominant trait but its dominance is partially modified by the A-locus, which basically means it can be present on an agouti mouse and not show depending on what else is going on. Few people have real umbrous in the US, as all of our umbrous comes directly from English imports. You can read more about umbrous on the Jax (no relation, haha) laboratories website, but I forgot where the exact link is since I lost all my bookmarks recently when my computer crashed...
Edit: Here is a photo of one of my heterozygous umbrous mice, WNT Langlia:
She is argente and the umbrous part of her is dove. Were she agouti, the dove part would be solid black, and were she black the umbrous part would help to make her darker all over. Despite good type (that you can't really see in this picture, she's an old lady), she could never be shown due (in part) to the umbrous.