It is a very great honour and privilege to be addressing you today. It has been a particularly interesting year politically at least as far as mathematics is concerned. Despite what seems in many ways a prosperous and optimistic situation in this country at the moment, mathematics is in a very serious state. Our subject is the key to living successfully in the modern technological world, but we have a grave shortage of mathematicians. In addition there are endless complaints from employers and the general public that young people leave school without the necessary mathematical skills to cope with everyday life.

Except for the circle, for which the true anomaly v is proportional to the time t, the position of a body in orbit about a central body at a given time is simplest to derive for a parabola. The classical determination of the time of flight on a parabolic trajectory is through the integration of the dynamic equations of motion. (See Appendix.)

In standard textbooks (e.g. [1, p. 877]) as well as in Encylopaedia Britannica [2, Vol. 13, p. 611], one finds variants of the following statement:

This statement is actually false, and can only be understood in the following sense: It is true only up to a linear approximation, and in fact, as soon as a mirror focuses parallel rays, it must be a parabolic mirror.

In the configuration, illustrated in Figure 1, ABC is a triangle with I1; I2, I3 the excentres opposite A, B, C respectively. The triangles I1BC, I2CA, I3AB are denoted by T1, T2, T3 respectively. Ok, Hk, are the circumcentre, orthocentre respectively of triangle Tk, k = 1, 2, 3. Lk is the Euler line of Tk, k = 1, 2, 3. Define the point Ek (t) on Lk by the equation

A topic of current interest is that of finding defining sets (preferably smallest defining sets) for various combinatorial structures such as latin squares and block designs, so it seemed interesting (though of no practical use) to try to solve the same problem for magic squares of small size. (The germ of this idea was suggested in the Gazette Note [1] of 1988 under the sub-heading ‘Squares from clues’.)

Smallest defining sets

A standard Sudoku square is a 9 × 9 latin square in which each of the nine 3 × 3 subsquares into which it can be separated contains each of the integers 1 to 9 exactly once.

A current problem is to complete such a square when only some of the cells have been filled. These cells are often called ‘givens’. (Such problems are currently (2005) published daily in British newspapers.) In more mathematical terms, the given filled cells constitute a defining set or uniquely completable set for the square if they lead to a unique completion of the square. If, after deletion of any one of these givens, the square can no longer be completed uniquely, the givens form a critical set. The investigation of critical sets for ‘ordinary’ latin squares is a topic of current mathematical interest. (See [1] for more details.)

The authors of the recent Note [1] exhibit an odd preference. They derive recurrence relations for the odd terms, un = F2n+1, n ≥ 0, in the sequence of Fibonacci numbers, Fn, defined by