P.R. using single member districts

[DEMOREP1 at aol.com]

Donald's comments about MMP prompted the following---

It is possible to have elections using single member districts to produce
approximate p.r. results in an N seat legislative body.

1. All partisan (P) and independent (I) candidates would get on the ballot by
an equal nominating petition. That is, the voters would determine the
candidates on the ballot (and not a party elite making up a party list). 
2. Voters would vote 1, 2, etc. for their choices.
3. The votes for each party in all districts would be totalled and the seats
for each party would be determined by a p.r. formula (which formula to use is
of some dispute).
4. If all candidates of a party (or an independent candidate) did not get
Total Votes in all districts/N (the ratio R), then the first choice votes for
such candidates would be transferred to the second (etc.) choice candidate
starting with the candidate getting the lowest number of votes.
5. The candidates of each party getting one or more seats would pick the seat
winners for such party using the number of votes that each candidate of such
party got.
6. Any independent getting R would be elected.
7. A candidate getting more than R could transfer the surplus to another
candidate in any district.

Shorthand- 
If a  P or I gets R, it, he or she is elected and can transfer surplus over
R.
If a P or I does not get R, then the lowest vote receiver loses and has votes
transferred to next choice.

More than 1 candidate in a district might be elected.  Some districts might
not directly elect any candidate. However, the votes in all districts would
be effective in electing someone (assuming enough second, etc. choices are
made).