this could actually could be quite interesting, so can some one explain to me, in depth, the old rules
The old rules are something like this:
1. Get the electorate of the country (each of England, Scotland, Wales & Northern Ireland is done separately) and divide by the existing number of seats. This gives the quota for the new constituencies.
1a. (Before 2005, Scotland had to have a minimum of 72 seats, and Wales had a minimum of 36 seats, so their quotas were smaller)
2. Draw up constituency boundaries which reflect local communities, ties, interests, and existing boundaries as much as practicably possible.
3. The big difference is that there were no minimum or maximum limits. The Boundary Commission imposed on itself a maximum variation of +/1 10,000, but even that was exceeded in some cases. The quota last time under the old rules (in the 1990s) was just over 70,000, so the Boundary Commission created constituencies between 60,000 ish and 80,000 ish.
4. The other big difference is that in most cases, the constituencies did not cross county boundaries. Before 1997, they didn't even cross the boundaries of London Boroughs. So, before 1997, each London borough had 2 or 3 or 4 constituencies. The boroughs which had only just enough electorate for 1.5 constituencies were rounded up to 2, and the ones with enough for 2.4 were also rounded down to 2. Therefore there was a much bigger variation in the electorates of London constituencies compared with the rest of the country.
For example:
Here is a county with 5 constituencies. But its current electorate is enough for 5.46 quotas.
Therefore it should have either 5 or 6 constituencies. But:
if it has 5 constituencies, each one will have an average electorate of (5.46/5) = 1.092 quotas (9.2% bigger than average);
if it has 6 constituencies, each one will have an average electorate of (5.46/6) = 0.910 quotas (9.0% smaller than average).
But a deviation of 9.0% is less than a deviation of 9.2%, so it is better to have 6 constituencies than 5.
(going off on a tangent)
In other words, the number of constituencies in a county was determined by the
geometric mean rather than the
arithmetic mean.
The effect of this was that any county or borough with an electorate of more than 2.40 quotas was rounded up to 3 (because 2.40/2 is 1.20 and 2.40/3 is 0.80);
any with an electorate of more than 3.428 was rounded up to 4;
any with an electorate of more than 4.444 was rounded up to 5;
and so on.
The way this worked was that it was easier to round up the numbers (and increase the number of constituencies) than to round them down. Even though this was not the original intention, that's why the number of MPs increased from 615 to 625 to 630 to 635 to 650 to 659.
After 1997, the Boundary Commission dealt with this situation by dealing with London boroughs in pairs, so that if two small boroughs were next to each other, they might have 3 constituencies between them (with one constituency straddling the boundary) instead of having 2 each. Or two medium-sized boroughs might have 5 between them instead of 3 each.
(going back from the tangent to the county)
So, if the county is increased from 5 constituencies to 6, the five existing constituencies would have been squidged about and a new sixth constituency would have been created in the middle.
They still used whole wards without splitting any.
The rules allowed for deviation from the normal rules if there were "exceptional" geographical circumstances, so Northumberland had 4 constituencies instead of 4 (Berwick and Hexham were two big ones), and in North Wales Ynys Mon stayed as a constituency on its own instead of being combined with part of the mainland.