Hall plane

From Wikipedia, the free encyclopedia
Jump to: navigation, search

In mathematics, a Hall plane is a non-Desarguesian projective plane constructed by Marshall Hall Jr. (1943).[1] There are examples of order p2n for every prime p and every positive integer n provided p2n > 4.[2]

Algebraic construction via Hall systems[edit]

The original construction of Hall planes was based on a Hall quasifield (also called a Hall system), H of order p2n for p a prime. The construction of the plane is the standard construction based on a quasifield (see Quasifield#Projective planes for the details.).

To build a Hall quasifield, start with a Galois field, F = GF(p^n) for p a prime and a quadratic irreducible polynomial f(x) = x^2 - rx - s over F. Extend H = F × F, a two-dimensional vector space over F, to a quasifield by defining a multiplication on the vectors by (a,b)\circ (c,d) = (ac -bd^{-1}f(c), ad - bc + br) when d \neq 0 and (a,b) \circ (c,0) = (ac, bc) otherwise.

Writing the elements of H in terms of a basis <1, λ>, that is, identifying (x,y) with x  +  λy as x and y vary over F, we can identify the elements of F as the ordered pairs (x, 0), i.e. x +  λ0. The properties of the defined multiplication which turn the right vector space H into a quasifield are:

  1. every element α of H not in F satisfies the quadratic equation f(α) =  0;
  2. F is in the kernel of H (meaning that (α  +  β)c  =  αc  +  βc, and (αβ)c  =  α(βc) for all α, β in H and all c in F); and
  3. every element of F commutes (multiplicatively) with all the elements of H.[3]


Another construction that produces Hall planes is obtained by applying derivation to Desarguesian planes.

A process, due to T. G. Ostrom, which replaces certain sets of lines in a projective plane by alternate sets in such a way that the new structure is still a projective plane is called derivation. We give the details of this process.[4] Start with a projective plane π of order n2 and designate one line \ell as its line at infinity. Let A be the affine plane \pi \setminus \ell. A set D of n + 1 points of \ell is called a derivation set if for every pair of distinct points X and Y of A which determine a line meeting \ell in a point of D, there is a Baer subplane containing X, Y and D (we say that such Baer subplanes belong to D.) Define a new affine plane D(A) as follows: The points of D(A) are the points of A. The lines of D(A) are the lines of π which do not meet \ell at a point of D (restricted to A) and the Baer subplanes that belong to D (restricted to A). D(A) is an affine plane of order n2 and it, or its projective completion, is called a derived plane.[5]


  1. All finite Hall planes of the same order are isomorphic.
  2. Hall planes are not self-dual.
  3. All finite Hall planes contain subplanes of order 2 (Fano subplanes).
  4. All finite Hall planes contain subplanes of order different from 2.
  5. Hall planes are André planes.

The smallest Hall plane (order 9)[edit]

The Hall plane of order 9 was actually found earlier by Veblen and Wedderburn in 1907. [6] There are four quasifields of order nine which can be used to construct the Hall plane of order nine. Three of these are Hall systems generated by the irreducible polynomials f(x) = x^2 + 1, g(x) = x^2 - x - 1 or h(x) = x^2 + x - 1. [7] The first of these produces an associative quasifield,[8] that is, a near-field, and it was in this context that the plane was discovered by Veblen and Wedderburn. This plane is often referred to as the nearfield plane of order nine.


  1. ^ Hall Jr. (1943)
  2. ^ Although the constructions will provide a projective plane of order 4, the unique such plane is Desarguesian and is generally not considered to be a Hall plane.
  3. ^ Hughes & Piper (1973, pg. 183)
  4. ^ Hughes & Piper (1973, pp. 202–218, Chapter X. Derivation)
  5. ^ Hughes & Piper (1973, pg. 203, Theorem 10.2)
  6. ^ Veblen & Wedderburn (1907)
  7. ^ Stevenson (1972, pp. 333–334)
  8. ^ Hughes & Piper (1973, pg. 186)