Alabama Senate redistricting.

[jimrtex]

This is the 2000 senate district plan drawn by the Democrats.  The first number for each district is the percentage black (total population), and the second the percentage deviation from the quota.



There are 8 majority minority districts, 3 in Birmingham, 4 in the Black Belt, including Montgomery, and one in Mobile.

The district in Montgomery was the one that duped the SCOTUS.  They saw it as a mostly rectangular district that followed county lines.   The city of Montgomery is in the northern part of the county, and the district actually envelopes the whiter portion of the city, with the enclave connected across the river to Elmore County.  The rural area in the southern portion of the county is simply a connector.

[jimrtex]

Though the deviation was kept within 5% limits, there is a bias with blacker districts tending to have smaller populations.  In addition, there was a tendency to push the deviation towards the limits, with the distribution totally inverted from what it would expected to be.

Absolute Deviation:

0%-1% 3 districts
1%-2% 4 districts
2%-3% 6 districts
3%-4% 6 districts
4%-5% 16 districts

The mean absolute deviation was 3.3% and the standard deviation 3.6%.

[jimrtex]

This is the 2000 districts, with the 2010 populations.  The first number is the black percentage (total population), and the second is the relative deviation.



The 8 majority black districts have lost considerable population share (equivalent to 1.24 senate districts collectively).  10 districts lost absolute population during the decade, including all 8 majority black districts.  The others were the purple district in the northwest along the Mississippi line, and the yellow Talladega-based district between Anniston and Montgomery.

There was considerable growth in the Huntsville area, particularly to the west.  I suspect most of the growth in the Limestone and Morgan based districts was in Madison County.  There was strong growth in the Birmingham suburbs, particularly towards the east which balanced the loss in the city core itself.  Generally, the suburban areas increased in share of the black population, but mainly from below 10% to the low teens.  The population loss in the black majority districts was due to largely blacks moving to better housing and job opportunities in the suburbs.  In the case of the Black Belt, there may be some die off, as 18 YO go away to college, then obtain jobs and bear children elsewhere.

[jimrtex]

This is the distribution of the 2000 districts based on the 2010 census.  Only 11 of the districts are within 5% of the ideal.  The correlation between percentage black (total population) and the relative deviation is 0.65 which is fairly strong.



I suspect that in the past, an effort was made to get the districts with the largest deviations barely within the 5% limits.  That could explain why the absolute relative deviation of 16 (of 35) districts was between 4% and 5% in 2000.  If we are attempting to maintain existing districts, we never fully correct the error.  A better approach would be like in Australia, where the objective is equality at 3.5 years into the distribution.  Growing areas would be deliberately underpopulated, and declining areas deliberately overpopulated.

More attention to the standard deviation and mean absolute value could also help.  For example, if a fair apportionment plan had a normal distribution of 2.5%, then 95% of districts would be within 5% of the quota, and the mean absolute value would just shy of 2% (1.995%).

The use of 1% as the maximum deviation by the Alabama legislature was totally reasonable in an effort to remove past embedded biases.

[jimrtex]

Clearly, there will need to be significant changes in the metropolitan areas.  But will there have to be interrerional adjustments as well?  We can group the senate districts as well.



Huntsville (5 greenish districts that include a portion of Madison County): 5.61 quotas.
Birmingham (8 blueish districts that include a portion of Jefferson County): 8.05 quotas.
Black Belt (7 greenish districts, including most of Tuscaloosa, Autauga, and Elmore counties) 6.74 quotas.

This separates three remaining regions:

Northwest (3 orangeish districts): 2.75 quotas.
Northeast (5 yellowish districts): 4.87 quotas.
Southern (7 orangeiish districts): 6.98 quotas.

The Huntsville region has a significant surplus, and assuming continued growth could support a 6th district by 2020.  But there is not an obvious region to lose a district.  So it would likely result in a continued contraction towards Huntsville.  Madison County has a population equivalent to 2.45 quotas but has only one district, plus 4 districts stretching into its neighbors (Madison County has a minority of the population in each of the 4 districts, such that a favored candidate might be from the other county, as smaller counties and cities tend to have a stronger sense of local community).

The Birmingham region has a tiny surplus, but most adjustments will be internal, as the three majority black districts must expand by around 25%, as the other 5 districts slide into the surrounding counties, particularly Shelby County.

The southern regions has about the right population to maintain 7 districts, but most significantly has no excess to share with the Black Belt.  The boundary between the two may be adjusted some, but the Black Belt will need to get its added population from the north.  There are major adjustments needed in Mobile and Baldwin counties, but they can be contained in that area.

The northeast and northwest regions, with a deficit of 0.13 and 0.25, respectively can take population directly from the Huntsville region.

The Black Belt needs to add 0.26 quotas (about 35,000) from the north, which must originate in the Huntsville region and be transmitted through the intervening regions (Birmingham, Northeast or Northwest).

A particularly awkward district is SD-30, which links Autauga County with Butler, Crenshaw, and Pike counties, with a narrow strip through Lowndes County.  Based on its number it is the process of being shifted from south of the Black Belt to north of the Black Belt (the districts to the east and west are SD-23, SD-25, and SD-26.  If a whole district were moved to the north, this would be the likely candidate.  Otherwise, it makes sense to try to shift it to the north of Montgomery.

[jimrtex]

As a first order adjustment we can assume that the Black Belt population can be augmented by the remainder of Tuscaloosa and Elmore counties, that in 2000 were in districts based further north.  If we also adjust the boundary of the Black Belt area southward, to include all of Choctaw, Clarke, Monroe, and Conecuh counties, we can swap Pike into the southern regions.   Elsewhere we align the regions on county boundaries.



Of the three counties in the southern part of SD-30, we can place Butler (41% BVAP) in the Selma-based SD-23, and Pike is now in the southern region.  We would still have Crenshaw (23% BVAP) to dispose of, and the Autauga portion of the district would have to be augmented to replace the loss of Butler and Pike.

The Birmingham region now has a deficit, due to the placement of the entirety of St. Clair in the northeast region.  The Huntsville region still has a sizable surplus.

We can make further adjustments.  We swap Henry (28% BVAP) for Crenshaw (23% BVAP), between the southern and Black Belt regions.   We move DeKalb from the Huntsville to Northeast region, St. Clair from the Northeast to Birmingham regions, and Chilton from the Birmingham region to the Black Belt region.



This results in the Huntsville, Northeast, and Jefferson regions having about the correct population for a whole number of districts.  The Black Belt region now has a surplus, but it is about the complement of the deficit for the northwest region.  Placing a portion of Tuscaloosa that is currently in a Walker-based district that also laps into Jefferson, into the northwest region balances the population of these two regions.

In the Huntsville region we can maintain the current map with one district based in Madison, and another district extending into another county, by sliding the districts about a bit.  Or we could place Marshall and Jackson into a district, and create a second district that is largely in Madison.

In the northeast, there will be major adjustments to accommodate the addition of DeKalb and the loss of St.Clair, but the large surplus in the Huntsville region supports the change.  Making gradual adjustments decade after decade  can lead to grossly distorted districts.

In the Birmingham area we were already faced with the need to expand the three black majority districts outward.  The loss of Chilton from SD-15 may actually reduce the adjustments, since the district had the second largest surplus in the state due to the growth in Shelby County.

Chilton (9% BVAP) is clearly not in the Black Belt, but its inclusion in the region is part of recognizing that the area can no longer support seven districts.  But the area with 6 districts can comfortably support four majority black districts largely comprised of whole counties.

[jimrtex]

This is the current senate map.  The black percentage is based on the 2008-2012 ACS (I could not find a version where the census bureau has tied census data to the legislative districts created after the census).



For comparison purposes, here are the 2000s senate districts with 2010 populations.



As anticipated, SD-30, the district that linked Autauga via a strip through Lowndes, to Butler, Crenshaw, and Pike, was moved to north of Montgomery.  Butler (41% BVAP) was shifted to SD-23, the Selma-based majority black district, while Pike (35% BVAP) was shifted to SD-31.  This left Crenshaw.  

Rather than maintaining a link through the rural Lowndes (72% BVAP), Crenshaw was linked to a district which, despite appearances, was mostly in Montgomery County.  It is ridiculous to think that you could take a district that was 75% black and entirely in Montgomery County, add 12% population to bring it back up to equal population, in a district that is 75% black and not have it maintain its core.

The plaintiffs had originally claimed that using the 1% threshold for equality was done with discriminatory intent, while the legislature had said it done so to overcome the past practice of deliberately underpopulating certain districts, and not correcting district populations to equality (remember the 2000 plan had 16 of 35 districts with a 4% to 5% threshold).  The were in essence targeting two 1% intervals with an 8% gap between them.

The SCOTUS was too embarrassed to address the claim, and said that it was a given that the legislature would strive for population equality.  The SCOTUS also appears to say that on remand the district court should focus on the majority black districts.

In the Birmingham area, the legislature maintained the basic configuration of the districts, while eliminating a 20% population deficit.  The alternative proposed by the plaintiffs maintains the black percentages in the three districts, while rearranging their configuration.  The district court may tweak the edges a bit, but they are unlikely to order the creation of four 45% BVAP districts, nor order bypassing of mixed areas to get to whiter areas.  And the legislature will be given the opportunity to act in any case.

The two western rural Black Belt districts SD-23 and SD-24 actually split fewer counties than in the old map.  Perhaps SD-24 could be adjusted a bit in the Tuscaloosa area, but nothing of significance.

Since the SCOTUS highlighted SD-26, the Montgomery black majority district, the district court may demand a change.   The legislature could begin by returning SD-26 to its 2000 version, and then moving adding area from SD-25 from the western arm of SD-25 (which is what they did in 2012).  This could clearly be justified on the basis of maintaining cores of existing districts, and communities of interest.

This leaves Crenshaw County isolated from SD-25.  But Crenshaw only has about 14,000 persons, about 10% of a senate district.  Possible solutions:

(1) Add it to SD-26.  The 2000 version of SD-26 would still be underpopulated, but less of a correction would be needed.

(2) Restore Crenshaw to SD-30, which is now based in Autauga, with the linkage through Lowndes.  SD-25 would take a bit more of Elmore.  Lowndes only has 11,000 persons total, so the connector likely only has a couple thousands persons, which can be balanced with a bit of Autauga.

(3) A super counterclockwise rotation.  Crenshaw would be shifted to SD-31 (red), part of Covington to SD-22 (green), shifting SD-23 (blue) or SD-24 (purple) southward simplifying the boundary, part of Pickens and Tuscaloosa to SD-21 (yellow).  The rotation could contnue around the north end of Birmingham or cut across to SD-14 (red).   This plan has the advantage of making the boundaries a bit cleaner, and knocking a few points off the BVAP%.  The senators from SD-31 and SD-22 are from Elba, Coffee County, and Bay Minette, Baldwin County, respectively, so their incumbency is unlikely to be challenged.   The downside of this plan is that more districts would be affected, which is a negative if special elections were triggered.  Alabama does not hold legislative elections until 2018, so they may wish to make only minor changes.  A court might decide that if a district changes by a small amount (say 10% or less of the population) that no special election was needed.

The final Black Belt district is along the Georgia border and east of Montgomery.  Its black percentage decreased from 56% to 51% from 2000 to 2010.   Race-neutral changes to the district risk reducing the black majority to a minority.  Arguably it had to have its black percentage increased to survive as a majority for another decade.  The changes to the district were the hook into Dothan, likely to pick up black voters, an expansion of the hook in Phenix City, also to pick up black voters, and a very intricate re-drawing of the boundary in Lee County, which appears to have the purpose of pulling Auburn out of the district (the 2000 boundary split the district).

An alteration might also provide a solution for Crenshaw County.  If the hook into Houston County were eliminated, the SD-28 could be expanded into Pike County, which is 35% black, and arguably would be maintaining a community of interest.  At the same time, Crenshaw could be shifted into SD-31, followed by adjustments to the north and east of Montgomery, pushing SD-28 out of Lee County.

The final majority black district, SD-33 is in Mobile.  The boundary is not particularly irregular.  If care was taken not to remove any territory from the 2000 district, the district could likely be configured largely on a community of interest basis which would tend to add blacker areas adjacent to the 2000 district.

The SCOTUS appears to have narrowed the case to the eight majority black senate districts.  They made saw the goblet rather than the face in Montgomery County.  This is something that the state can concede and readily "correct" by returning SD-26 to its 2000 configuration, and then increasing its population, by territory that was already shifted into the district.

The district court might not or might not find fault with the 3 Birmingham districts and the one Mobile district.  But it is quite unlikely that the court would order adding areas that were not adjacent to the 2000 districts.

Meanwhile, no attention has been paid to the main effect of the redistricting effort.  Before redistricting, nine districts had a population that was between 24-36% black.  After redistricting, this had dropped 9%.  The three districts over 30%, declined an average of 14%.  The gap between the 8th blackest and 9th blackest district widened from 36% to 51% (gap of 15%) to 26% to 59% (gap of 33%).

[jimrtex]

In one of their post-remand briefs, the plaintiffs argue that by adopting the 1% deviation criteria, the legislature had violated the Alabama Constitution's provisions that senate districts not split counties, since such a narrow band caused additional counties to be divided.  Since the SCOTUS has permitted a deviation of 5% to comply with constitutional requirements such as respecting political boundaries, the plaintiffs claimed that the legislature's use of a 1% violated OMOV.

Section 200 of the Alabama Constitution states:

"... divide the state into as many senatorial districts as there are senators, which districts shall be as nearly equal to each other in the number of inhabitants as may be, and each shall be entitled to one senator, and no more; and such districts, when formed, shall not be changed until the next apportioning session of the legislature, after the next decennial census of the United States shall have been taken; provided, that counties created after the next preceding apportioning session of the legislature may be attached to senatorial districts. No county shall be divided between two districts, and no district shall be made up of two or more counties not contiguous to each other."

So the requirements are:

(1) Population equality to the extent possible;
(2) No splitting of counties between districts;
(3) Multi-county districts must be formed from contiguous counties.

The constitution has not been modified since Reynolds v Sims, which was based on the malapportionment of the Alabama legislature.  In particular, it was the 2nd provision which prevents division of larger counties that caused the SCOTUS to rule in that case (plus the fact that Alabama had never redistricted under terms of the 1901 constitution).

If the Alabama Constitution were followed, the senate map might look like this:



Eight counties, Jefferson (4.822 quotas), Mobile (3.024), Madison (2.452), Montgomery   (1.680),
Shelby (1.429), Tuscaloosa (1.425), Baldwin (1.335), and Lee (1.027) have a population of more than one quota, and thus would have their own senate seat.

Collectively, they have 17.193 quotas.  Distributing the population of the remainder of the state among the 27 remaining districts, results in a new quota that is 66% of the original quota.  An additional 6 counties exceed this reduced quota, Morgan (0.875), Calhoun (0.868), Etowah (0.765), Houston (0.744), Marshall    (0.681), and Lauderdale (0.679), and are also given their own senate seat.

The 14 largest counties have a population equivalent to 21.804.  Distributing the population of the remaining 53 counties among the 21 remaining districts, gives a new quota that is 62.8% of the original quota.  This is the ideal size of the remaining 21 counties.

Under the above map, an additional 6 counties have a single-county district, because their population is near the ideal size, or there are very small adjacent counties to be added to them.  These counties are Limestone, Cullman, St. Clair, Talladega, and Elmore.

Two relatively large two-county districts are DeKalb-Jackson and Walker-Blount.  They were joined to avoid a district population with less than half of the original quota.  While they are large relative to the other small county districts, they are still less than the original quota.

Three districts are majority black.  Based on their largest county they are Dallas (63% black), Montgomery (52%), and Russell (55%).

While the map would not pass muster under OMOV, they could conceivably be used with a weighted voting scheme.  The very large Jefferson, Mobile, and Madison districts would likely be a problem.

[jimrtex]

This is the order of the district court on remand of Alabama Legislative Black Caucus v State of Alabama.

It says that the court will only consider whether race predominated in the drawing of the 28  majority black House districts, and the 8 majority black Senate districts.

The SCOTUS said that the district court might reconsider other issues, which the SCOTUS did not address.  The district court said it will not reconsider:

(a) Whether the use of the 1% deviation violated One Man, One Vote;
(b) Whether it was unconstitutional partisan gerrymandering.
(c) Whether the interaction between the districts and the local legislative system₁ violates equal protection.
(d) Whether there were violations of Section 2 of the VRA.

So the lower court decision is going to be very focused.  At most some very minor changes to senate districts will be required.

1 County governments in Alabama are particularly weak, and local legislation is performed by the legislature.  As might be expected, the legislature is typically deferential to the legislatures from the county.  And in Alabama, there are formal committees with subject matter authority for local legislation for the larger counties comprised of the members that represent parts of the county.  In Jefferson County, there are three majority black senate districts in the county, and parts of five others.

[jimrtex]

This map is based on a more liberal interpretation of the Alabama Constitution, or would be an attempt to harmonize the Alabama Constitution and the US Constitution.

In 1880, Jefferson County was the 20th most populous county in Alabama.  By 1890 it was the most populous.  In 1901 when the current consitution was written, it would have been the only county clearly entitled to have more than one senator.  Mobile and Montgomery had a population somewhat more than a quota, but hardly enough that would challenge representational ideals of the time.

Jefferson County grew significantly during the first half of the century.  By the time of Reynolds v Sims, Jefferson would have been entitled to almost 7 senators (since 1960, Jefferson has gained less than 4% in population).

The constitution requires election of only one senator per district, and the splitting of a county between districts.  The election of only one senator from a single-county district is what causes the OMOV violation.   So the Alabama Supreme Court could overturn that restriction.  From a 1901 perspective when the constitution was written, the concern might well have been more about a county being chopped into two districts.  Two halves do not make one whole, particularly if the senators are from other counties.

The Alabama Supreme Court overturns the limit of one senator per district, and the associated requirement that the number of districts be the same as the number of senators.  It maintains the restriction on dividing counties and that districts be composed of contiguous counties.  It interprets the requirement of district population equality to mean average population per senator from the district equality.

The following plan is based on apportioning the 35 senators among the 8 counties with a population greater than one quota, and the aggregate population of the other 59 counties.

The large counties are collectively entitled to 17.807 senators, and are apportioned 18 senators using a list method utilizing the harmonic mean.  This is done with the goal of keeping the population per senator as near the average as possible.  It results in the smaller counties typically having a population greater than the quota.  This is exacerbated by some districts in the northern part of the state having a population below a quota (eg Morgan, DeKalb-Jackson, and Calhoun-Cleburne-Clay).

Weighted voting would be quite workable in this case.  Initially, the senators from a multimember district might have been elected at large, later switching to election from subdistricts.



21 (sub)districts have a population within 10% of the ideal;
6 are within 20%;
7 are within 30%;
1 (Baldwin) is within 40%.

The standard deviation is 12.0%.

[jimrtex]

As part of their remand brief the plaintiffs offered this map, which reduce the number of county splits, while keeping districts within 5%.   It must be remembered that one of the plaintiffs drew the structure of the current map in the 1990s litigation, and then helped underpopulate the black majority districts in 2000 (when 16 of 35 districts had deviations between 4% and 5%).   When the Republicans updated the districts in 2010, they reduced the maximum deviation, rather than risk complaints of racial gerrymandering if some black majority districts had much surplus.



The plaintiffs now argue that the 1% threshold had discriminatory intent.  The district court dismissed that complaint.   The SCOTUS ignored the thrust of that complaint and said that it would simply be a matter of course to equalize population.  On remand, the plaintiffs offered a 5% plan which they said (better) respected the provision against county splitting in the Alabama constitution.  The district court has now dismissed the 1% threshold complaint on remand.

The map below is a stylized version of the map proposed by the plaintiffs.

n le

It splits 8 counties with a population less than a quota in population, and splits two larger counties non-optimally.  An optimal split is when a county with population in excess of quota, is given as many whole districts within the county, with the excess placed in a single district.  Thus Montgomery, Tuscaloosa, and Baldwin are optimally split since they have one district in the county, and the remainder in a single district that extends into adjacent counties.  In addition, Jefferson (5), Mobile (3), and Lee(1) have close to a whole number of quotas, and have that many districts.

Shelby has a population equivalent to 1.429 quotas, but is split between two districts, rather than having one district of its own.  Madison has a population equivalent to 2.452 quotas, but has one district of its own, and parts of two others, rather than two of its own.

The plan also has a 5-district area east of Birmingham, extending from Marshall to Bibb and Chambers counties.  In total, only 6 (of 35) districts comply with the Alabama constitutional requirement that no counties be split.

[jimrtex]

Given the proposal by the plaintiffs, I wondered if I could do better.  And I could.



This map has a maximum deviation of 5%.  It divides 5 smaller counties, rather than 8, and divides all the larger counties optimally with the maximum number of whole districts within those larger counties.

Only two districts have an absolute deviation greater than 4%, and 11 have an absolute deviation less than 1% (this assumes all county splits will have equal population).  The mean absolute deviation is 2.0% and standard deviation is 2.4%.

It is likely that there could be 6 majority black districts, and 4 influence districts.

[muon2]

Have you made graphs of deviation vs BVAP for the plaintiff's map and your offering? I like your plan except for the Baldwin-Washington connection. However I see that the plaintiff's map has that, too, and essentially it is in the current map as well.

[jimrtex]

Have you made graphs of deviation vs BVAP for the plaintiff's map and your offering?

I didn't draw individual districts, so it makes it difficult to compare.  Looking at the districts that split counties, I didn't see much variation based on BVAP, for example in the Birmingham districts, or the Mobile districts.  There is some variation in that blob east of Birmingham, with Shelby districts having a deviation above 4%, and BVAP below 10%, and the district on Georgia border having a deviation of 1.4% and 25% BVAP.

It was that area that attracted my attention in the first place.  In the rural areas of the state, the counties have around 8% to 15% or so of a quota, and it is pretty easy to draw whole county districts.  But you get into some pretty substantial populations in the Gadsden, Anniston, Talledega area.   But the boundaries were drawn so squiggly that I was curious as to why.  And I still don't know.  It might make sense if you were trying to get some districts up past 30% BVAP, but not to differentiate into 18% and 8% districts.  The Talladega district in the Plaintiff's plan is 25% BVAP, and made of whole counties.

In the 2001 map, the Talladega district had 35% BVAP in 2010.  The Alabama River splits into the Coosa and Tallapoosa rivers past Montgomery, and there is a tapering out from the Black Belt as you head north.  But the area is now somewhat separate because of the growth of the Montgomery suburbs in Elmore and Autauga counties.  Talladega County is now split among four districts.  One of the main effects of the 2012 redistricting was to reduce the BVAP% of districts in the 25% to 35% range.  For the most part this was not done by pushing the percentage higher in the eight majority black districts.  They had already been drawn in a way that produced a large variation from neighboring districts.

The plaintiffs remand plan was primarily a litigation tool to demonstrate that you really didn't need to go to a 1% deviation limit.  But it was also an effort to do a do-over, to make some changes that had nothing to do with using whole counties.

If you are going to emphasize whole counties as a fundamental principle, second only to population equality, you need to stick to it.

I like your plan except for the Baldwin-Washington connection. However I see that the plaintiff's map has that, too, and essentially it is in the current map as well.

The Baldwin-Washington district is SD-22, the last Birmingham-area district is SD-20, and the Tuscaloosa district is SD-21.  SD-24 and SD-23 are to the north, SD-25 and SD-26 are in Montgomery, and SD-27 and SD-28 are along the Georgia border.   SD-29 is in Dothan and the districts are numbered through SD-35 in the southern part of Mobile County.

So it appears that SD-24 and SD-23 have been pushed westward and expanded southward, forcing SD-22 to the south and SD-21 to the north (and SD-24 has ended up west of SD-23).  Meanwhile, SD-30 was shifted through the Black Belt so it is now north of Montgomery, which has enabled SD-22 to also extend to the east.

So Washington is more like the last outpost of the traditional district.  Mobile is growing a bit slower than the state, so that in the 2001 map, there was a portion of SD-22 that extended southward along the Alabama-Mobile river to the city of Mobile.   Meanwhile Baldwin has been growing faster than the state, and is now contributing about 1/3 of the district, and the current senator is from Bay Minette.  The current district with the very irregular boundary across Choctaw, Clarke, Washington, Monroe, and Conecuh appears to be an attempt to put the majority black areas in the districts to the north, but keep areas with significant black population (perhaps 20-30%) in the district.

[Torie]

Does any of this exercise have anything to do with trying to predict what the lower court will do with the map on remand from SCOTUS?

[jimrtex]

1960s

Following Reynolds v Sims, the legislature redistricted.  It had also redistricted during the litigation, but limited by the Alabama constitution (maximum one senator per county; minimum one representative per county) could not provide much relief.  The initial litigation was primarily by people in Jefferson County which at the time was entitled on a population basis to 6.8 senators (and had 1), and 20.4 representatives and had 21.  At the same time, Lowndes which had a population equivalent to 0.17 senators and 0.51 representatives, had one senator and two representatives.

After the district court had made its initial rulings, Jefferson had its number of representatives increased to 17.  The plaintiffs did not have enough money to appeal the decision.  But a (white) probate judge, Bernard Reynolds from Dallas County, appealed the decision, concerned that the (white) voters in Dallas County being swamped by black voters from Lowndes County.  This enabled the original plaintiffs to cross-appeal and get a much more sweeping decision from the SCOTUS, and gave the case its Reynolds name.

The senate apportionment in the Alabama constitution that fundamentally conflicted with OMOV was the requirement that each senate district elect one senator.  If a senate district could elect more than one senator, then county lines could be preserved.  The legislature reapportioned giving multiple senators to Jefferson (6.80 quotas, 7 senators), Mobile (3.37, 3); Montgomery (1.81, 2).  The other counties with more than one quota at the time were Madison (1.26), Tuscaloosa (1.17), Etowah (1.04), and Calhoun (1.03).   The district court that considered the senate plan in Sims v Baggett was somewhat concerned that Madison had only one senator, and a 26% deviation, but felt it could be mitigated by apportioning additional representatives to the county.  The district court approved the senate plan, and noted that the the largest district was only 1.45 times the population of the smallest.  It had been about 33 to 1 previously, so this was a significant improvement.

The court at the same time rejected the House plan, both because it had combined smaller counties in multi-member districts.  For example, Elmore, Macon, and Tallapoosa had 3 representatives, though the population of the three was not grossly dissimilar, ranging from 26K to 35K, with the apparent intent of preventing election of a black representative from Macon.  Mobile, though entitled to slightly more than 10 representatives was granted 9. 

The court also said Madison should have a 5th representative to compensate for being underrepresented in the senate.  Madison was entitled to 1.257 senators, and 3.772 representatives.  If considered independently, this would mean 1 senator (25.7% deviation) and 4 representatives (-5.7% deviation).  The courts logic appears to give Madison 7.544 representation units (out of a total of 210), round this to 8 representation units, and give Madison one senator worth 3 representation units, and five representatives worth one representation unit each, for a total of 8 representation units.  This could be considered a form of weighted representation in the legislature via a composite of senators and representatives.

[jimrtex]

Does any of this exercise have anything to do with trying to predict what the lower court will do with the map on remand from SCOTUS?

Yes.