Ive had a hack around with the distribution of catters in the UK. What is immediately apparent from a preliminary inspection of the frequencies is that while the available observations do follow predicted norms, some error has crept in to the results.

The specific outlier in question is in the Yorkshire cell, and the extremity of the degree of deviation can only realistically be attributed to experimental error. Therefore in accordance with standard scientific protocols, the Yorkshire cell has been disregarded in all subsequent analyses.

So, taking into account a distribution of UK catters (standardised into density rather than absolute numbers) and ignoring the outliers in the usual way, a set of travel time polygons were fabricated and superimposed on a node map of a trial population of centroids. The relative locations of the nodes were then manipulated using a Fubar transform to generate a best-fit solution.

Therefore, I can state with the usual professional certainty that, based on travel time, catter density, the absence of Yorkshire from the equation and a 'number of other factors' (ahem) the best fit result for the problem lies on the A5 just as it crosses the Bucks/Northants border.

Thank you for your time, ladies and gentlemen. Now if there are any questions?