[Illustration: FIG. 16.--Maze at Hatfield House, Herts.]

[Illustration: FIG. 17.--Maze formerly at South Kensington.]

[Illustration: FIG. 18.--A German Maze.]

The maze at Hatfield House (Fig. 16), the seat of the Marquis of
Salisbury, like so many labyrinths, is not difficult on paper; but both
this and the Hampton Court Maze may prove very puzzling to actually
thread without knowing the plan. One reason is that one is so apt to go
down the same blind alleys over and over again, if one proceeds without
method. The maze planned by the desire of the Prince Consort for the
Royal Horticultural Society's Gardens at South Kensington was allowed to
go to ruin, and was then destroyed--no great loss, for it was a feeble
thing. It will be seen that there were three entrances from the outside
(Fig. 17), but the way to the centre is very easy to discover. I include
a German maze that is curious, but not difficult to thread on paper
(Fig. 18). The example of a labyrinth formerly existing at Pimperne, in
Dorset, is in a class by itself (Fig. 19). It was formed of small ridges
about a foot high, and covered nearly an acre of ground; but it was,
unfortunately, ploughed up in 1730.

[Illustration: FIG. 19.--Maze at Pimperne, Dorset.]

We will now pass to the interesting subject of how to thread any maze.
While being necessarily brief, I will try to make the matter clear to
readers who have no knowledge of mathematics. And first of all we will
assume that we are trying to enter a maze (that is, get to the "centre")
of which we have no plan and about which we know nothing. The first rule
is this: If a maze has no parts of its hedges detached from the rest,