The minor pieces are bishop and knight (and the major
rook and queen). They are of roughly equal value (3 pawns, we
often say), but have such different powers of movement that they have
very different uses in the endgame.
In the struggle of bishop against pawns compared with
pawns, the superiority of the bishop is plain, the knight being more
easily overstretched or herded away. However, there are
some neat mates available to the knight when the side with the king has
a rook's pawn, which should be known. And there is a quirk of
chess, that a rook's pawn which will queen on the opposite colour to
its supporting bishop, cannot be forced through against a defending
king. But, on the whole, we would sooner go into an
ending with a bishop than a knight.
Knight endings remind us of King and Pawn endings,
knights are similarly slow across the board, and are subject to zugzwang. They can
also run out of manoeuvring space at the edge
of the board, and hate to chase rook's pawns. When knight battles
knight, it is often hard to calculate how many moves it will take a
knight to eat up a group of pawns, because a knight takes three moves
before it can attack the next square on a rank, and then the pawn might
move... With a bishop, if the pawns cannot escape, it's very easy
to count your way through a sequence.
Endings with opposite coloured bishops are notoriously
drawish, because the defender can so often set up a blockade.
Endings with bishops of the same colour have some
features. You are usually better off with a pawn at the edge of
the board, so it's easier to squeeze the defending bishop off the
defence of the queening square when the defending king is
The battle of bishop(s) against knight(s) in the endgame is worth studying. When there is play on each side of the board, the bishop is superior -- so much so that you can play for a win based just on that advantage. However, when the play is all on one wing, the capacity of the knight to control squares of each colour can make it the preferable piece.
If you share Simon's affection for studies, a very nice
collection of simple studies is Chernev's Practical Chess Endings, and this
includes dozens of minor piece endings.
However, if you actually want a guide to practical chess endings, then
I think Paul Keres' title of the same name is the best. (I
haven't seen Silman's or Dvoretsky's manuals.)