Yesterday was the School’s third Research Day, a successful and enjoyable event involving contribution from all divisions. Hopefully the event is now self-sustaining.

Short summaries of a few of the talks follow.

The first two speakers both had “automata” in their titles and both apologised for not talking about them due to shortness of time. Alan Hood told us about avalanche models of solar flares; these have been done using cellular automata, which don’t really take the physics into account. He and his colleagues have produced the first demonstration based on the differential equations of magnetohydrodynamics.

Then Tom Bourne spoke about regular languages. These are obtained from some basic building blocks (the empty set, the set containing the empty word, and the set containing a single one-letter word) by closing under union, concatenation, and the “star operation”. In these terms, the “star-height” is a measure of the complexity of a language. Noting that the set of regular languages is closed under complementation, he defined a “modified star-height” which allows the use of complementation in the construction. Now not a single regular language with modified star-height greater than one is known; do any exist?

Isobel Falconer told us about Maxwell’s encounter with the inverse square law of electrostatic attraction. It was basic to his main work; towards the end of his life he turned his attention to testing it experimentally. The inverse square law implies that there is no charge inside a closed conductor; this can be tested experimentally, but does the converse hold? Maxwell’s demonstration of this was flawed since the “no charge inside” principle implies the inverse square law if it holds for all possible radii of the conducting sphere, while he only tested one radius.

Helen Burgess talked about transfer of energy to larger scales (inverse cascades) in turbulent flow with vorticity, and found universal phenomena (in particular, three different scaling regimes) which seem to apply in completely different phenomena also.

From Patrick Antolin’s talk, I learned something I didn’t know: it rains on the sun! This puts the song “The sun has got his hat on” in an entirely new light!

Jonathan Fraser and his student have a remarkable result. Erdős and Turán posed the problem: if *X* is a set of natural numbers such that the sum of reciprocals of its members diverges, does *X* necessarily contain arbitrarily long arithmetic progressions? (The special case of the primes was solved fairly recently by Green and Tao, and was a big breakthrough.) The problem appears inaccessible, but they have proved an approximate version: such a set contains subsets which are arbitrarily close to long arithmetic progressions, in a suitable sense. Indeed, they prove this under the weaker assumption that *X* has Assouad dimension 1.

Negative feedback loops in gene regulation can produce oscillatory behaviour. The mechanism was not clear until Mark Chaplain showed that diffusion was a necessary part of the process. Cicely Macnamara told us about further investigations of this process, which can generate segmentation of bodies in embryonic development.

Finally, Alex Craik told us about William Welwood, St Andrews’ first professor of mathematics. He lived in difficult times, in the troubles between Episcopalians and Presbyterians following the Scottish reformation, and indeed was stabbed more than once and later forced to resign his chair. His one known output is a scheme for removing water from coal mines, which he proposed to do with a siphon, although he admitted that tests of the principle had been unsuccessful. (This was in the 16th century, before the work of Galileo and Torricelli; atmospheric pressure was not understood then.) Purely by chance, we had been reading about Culross, a village in western Fife which had a coal mine in the 16th century extending under the Firth of Forth, which eventually closed because of the water that leaked in; Welwood’s invention would not have helped in this case!