From VTDigger.org:
The two-member Essex-Orleans district became a one-member shed Essex County and some towns in Orleans County to the Caledonia district, which lost its portion of Orange County which combined with Windsor County to form 2 two-member districts, and other changes are described in the article although it's not clear exactly what changes are changes from the plan the LAB had adopted last week before they realized it had one two many Senators. You can view the plan and the population and deviation of each district here. (The Deviation column is the districts population minus the ideal district population of a district electing that number of Senators, but the "% Dev" column is the percentage that that district's population per Senator is over the statewide population per Senator. In other words, it's correct, although I thought it wasn't at first.) I verified the figures on a spreadsheet I made back when I started this thread.
I think they probably still had the Windsor 3-seater. This is the problem when you mix up apportionment and districting.
Franklin 2.29, Lamoille 1.17, Washington 2.85, Addison 1.77, Rutland 2.96, Windsor 2.72, Bennington 1.78, and Windham 2.13, have grossly enough population for a whole number of senators (18 total vs. 17.67 ideal).
But if you place Grand Isle with Franklin (you wouldn't want to stick it in the Chittendon super district) you then have 16 total vs 15.38 ideal.
Franklin-Grand Isle 2.62, Orleans-Essex 1.61, Caledonia 1.50, and Orange 1.39 are pretty far removed from a whole number of senators. But if you are just apportioning, you apportion 8 seats (3,2,2 and 1) vs 7.12 ideal.
Finally you have Chittendon with 7.51 ideal and even if you truncate to 7.00 you have 31 senators.
If you decide to split Chittendon, you end up with a 3-2-2 or a 2-2-2-1 split. The only other one-seaters would be Lamoille and Orange. And if you had a 3-seater, you might be tempted to put Burlington in it. But Burlington is right at 2 seats (2.03).
So you add Grand Isle to Chittendon so that there can be 8 districts (7.84 ideal), split 2-2-2-2 including a Burlington district. This restores Franklin to a 2-seater. The whole-unit counties then have 18 senators (17.67 ideal).
Orleans-Exeter, Caledonia, and Orange have 5 senators (4.50 ideal). You convince yourself that this is OK since Orange is significantly past enough for one senator, and you can get some extra people from Franklin and Lamoille. If you add these two to the NE counties. You have 8 senators (7.96 ideal).
But this leaves the remaining whole-unit counties at 15 senators (14.21 ideal). The excess in Franklin and Lamoille is being used two places. To shore up the population in the neighbors to the east; and to rationalize the apportionment of senators to the whole-unit counties. But 5 of the 6 whole-units counties in the south are under-populated (Washington 2.85, Addison 1.77, Rutland 2.96, Bennington 1.78, Windsor 2.72). The exception is Windham with a modest surplus 2.13.
So in total, this southern region had 15 senators for and ideal population of 14.21. The extra (31st) senator was spread over 6 counties, which was why they couldn't locate him.
Alternatively, the original plan had Orleans-Essex (2), Caledonia (2), Orange (1), and Windsor (3). Windsor would have had a deficit of 9.1%. Caledonia and Orange would actually be paired with a total deficit of 3.7% (in 2001, a fairly large chunk of Orange was placed in the Caledonia district.
The revised plan has Orleans (1), Caledonia-Essex (2), Orange (2), and Windsor (2) for one less district (8 to 7). Since this northeastern area ends up with a collective surplus of 0.29 this suggests most of the patching was done here.
The final plan shifts Charlotte (Chittendon), Elmore (Lamoille), and all of Orange to the south, while only adding one senator. This is equivalent to 1.62 senators, bringing the southern total to 15.83 ideal and 16 actual, and eliminates the hidden extra senator.
That's true (the overall deviation in 2002 was 14.73%, if defined as the difference in the largest population per Senator to the smallest as a percentage of the ideal district population which seems to be what federal courts use (rather than as a percentage of whichever district's population would yeild a larger figure which would always be the smaller district, which would also always yeild a larger figure than the ideal district population being the denominator unless there was no deviation), ...
Of course they should use the maximum absolute deviation from the ideal vs. 5%. They are trying to hit the center of the target, with some area around it counting as a score. The current court practice is to shoot the arrows and then shift the target to match.
That's true, but how has Vermont gotten away with such deviations? Maine had a House District in Limestone that had no more than 27.?% of the ideal district population (if all the census 2000 population in Limestone was in that district) from 2002 for 2004, but that was due to a delay in redistricting (and Loring Air Force base in Limestone having closed in the 90's) rather than a deviation as large as Vermont's when the districts were drawn. Perhaps it's all for the same reason: no one bothered to challege it. A 12 or 13 percent deviation in Maine's 1983 House district plan was upheld by the state Supreme Court court, but from a newspaper article a plan recommended by the challegers had an even larger deviation.
For the House of Representatives it appears that Vermont considers +/- 10% (twice the conventional federal standard) as normative.
+/- 5% is considered safe harbor, with anything larger having to be justified by the State. In some cases this has been done. Wyoming at one time required each county to have one representative (or senator) and it turned out that it wasn't too horrible to do so. Hawaii also apportions legislators by island group, though they have sometimes ignored the provision and created canoe districts that include a portion of Oahu with a portion of Hawaii or Maui. In 2001, they danced around whether or not to include the military population in order to rationalize eliminating districts, when they probably could have done so regardless.
Vermont can probably justify larger deviations for House districts, since they respect town lines. I think I'd be more comfortable if they first did an apportionment by county (or group) in the case of Grand Isle-Franklin and Essex-Caledonia, and then districted those individually.
But if they start dividing towns more extensively, it might be harder to rationalize.
I think Vermont would have a hard time justifying this most recent proposal for the Senate. They can't claim that they are apportioning on the basis of counties, when they shift so many towns (9 of 23 in Windsor, and 24 of 255 overall), and there is nothing in the constitution that requires them to do so. In addition, with some minor adjustments they could get within 5% (and eliminate two split counties). Too often it appears that the goal is to get districts within acceptable ranges rather that to the targeted population.
So if challenged, they would argue that they respect counties, but not really, and that there is no constitutional requirement, and that they shift towns to balance population but don't do a very good job at it.
For example, Bennington starts out at 1.78 and Windham at 2.13 so they shift some population to get Bennington somwhat closer (11% to 6%). What is truly odd is that Readsboro and Searsburg are moved from Bennington to Windham. If you put these two towns back in the Bennington district, Bennington is within the 5% range, and the two districts are more equal in population (1.93 and 1.98).
Caledonia-Essex started out at 1.80 and ended up at 2.16, for the worst deviation of any district. But if you move Wolcott (Lamoille) to Washington, then Caledonia-Essex is reduced to 2.08; and Washington is up to 2.98. You also eliminate the 3-way split of Lamoille. Alternatively, shift Stowe form Lamoille to Washington (and keep Elmore and Wolcott in Lamoille), and Lamoille would also be within 5%.
Then shift Westford to Grand Isle-Chittendon and St.George to Chittendon East, and all districts will be within 5%. The number of towns shifted would be reduced from 24 to 21 and county splits from 9 to 7.
If Vermont wanted to go to consistent 2-member Senate districts, they could divide the state into the following regions: East (Essex, Caledonia, Orange, Windsor, Windham) 8.04 senators;
West (Addison, Rutland, Bennington) 6.51; North (Grand Isle, Franklin, Orleans, Lamoille, Washington 7.95); and Chittendon 7.51.
Place West and Chittendon together, for 16.02 with one district taking in the more rural eastern part of Chittendon together with northern Addison, then the other districts will be fairly similar to what is proposed, while eliminating the the remaining two 3-seat districts and two 1-seat districts.
Parts of Rutland would be added to Addison, and parts of Washington added to Lamoille. The two-seat Rutland district would be centered on the city of Rutland, while that in Washington would be centered on Barre-Montpelier. The final one-seater in Orleans would become a two-seater with all the county along with the more rural part of Franklin.