|
First, "directly similar" means similar in the same rotational sense, as
opposed to "inversely similar", which means similar in the opposite rotational
sense.
The FTDSF says if the lines connecting the corresponding vertices of two directly similar polygons are divided in equal ratios, then the resulting polygon is directly similar to the given two polygons.
Expressed in set notation, the theorem is: Let F0 and F1
denote two directly similar figures in the plane, where P1
F1 corresponds to P0
F0 under the given similarity. Let r
(0,1), and define Fr = {(1-r)P0+rP1 : P0
F0,
P1
F1}. Then Fr is also directly similar to F0.
By considering the vertices of the first square in order BCDA, and the second in order PQRS, with r=1/2, then the midpoints of BP, CQ, DR, and AS are the vertices of the red figure, which by FTDSF is similar to the other squares.
proof based on post by Tom L, . . . . . . needs to be fixed to change "k" to "r", and add a diagram.
('Euclidean' proof)
It's fairly clear that the result will follow from the following:
Let AB, CD be segments and P be a point ('centre of spiral similarity') such
that ABP, CDP are similar. Then for E,F with kAE=(1-k)EC, kBF=(1-k)FD, EFP is
similar to the first two triangles.
Proof (drawing pictures will make it obvious what I'm doing):
Firstly, it's a nice trivial angle chase (or a 'standard result on spiral
similarities' if you like) to show that ACP and BDP are similar.
Now, our funny k-ratio condition really comes down to the fact that PBF is
similar to PAE (since PAC similar to PBD and E,F are points on the sides AC,BD
with the same ratio).
But now we are basically home... as <FPE = <FPA-<EPA = <FPA-<FPB = <BPA, and
FP/EP = BP/AP. So PAB is similar to PEF.
I'll leave the details on why this implies the whole result to you guys...
Cut-the-knot: Similar Triangles, which explains the Fundamental Theorem of Directly Similar Figures
Summary of geometrical theorems
The Finsler-Hadwiger Theorem -- Squares ABCD and PQRS joined at a corner D=P; the figure formed by joining the midpoints of AC, CQ, QS, and SA is a square. This is a special case of the FTDSF.
The webmaster and author of this Math Help site is Graeme McRae.