Take lane "multiplication" in SONAR Platinum Manchester?

I just had something strange happen in the Manchester version of SONAR Platinum.  I have no clue how it happened, so can't provide a recipe, but the practical effect was that the number of take lanes on one track got multiplied by 4 (and I have no clue if the even multiple is meaningful or just coincidence).  Here's what I know I did.  I am not sure, though, where along the line the extra lanes crept in:
 
1) Tracked multiple takes (using sound-on-sound, store takes in single tracks, and create new lanes on overlap) for two separate sections of the song in a single track.  After doing each section, I cut up the takes into phrase-length clips, deleted extra space, and applied trimming.  After recording the second section, I also had to drag all the clips from the first section up to the top and manually delete the now blank/extra take lanes.  I also manually isolated clips of one of the take lanes so it could serve as a reference for additional part tracking.  (I *really* wish SONAR would reuse take lanes if there is nothing in the new section of takes being recorded -- and I'd rather have most recent takes on the bottom, not on the top, but that is personal preference; what I don't want is a zillion lanes when I only need a as many as the section of the song with the most takes, or that plus any overlaps between sections, if applicable.)
 
2) Tracked the second section of the song (as above in all ways) in a separate take (additional background vocal part if anyone's curious), also doing the cutting up and applying trimming and isolating a lane for reference.  (FWIW, I do this to make comping easier later on. I prefer clip-level muting of phrases to having to drag a mouse around because it makes the auditioning process much easier when using parts from multiple lanes.)  There were exactly 9 takes in this part.
3) Tracked a third section of the song in the first track, and did the same drill with cutting things up into clips and applying trimming, as well as moving the earlier clips up to the top and deleting extra/blank takes as before to consolidate things, as well as the isolating clips from one of the lanes for reference.
 
4) Went back to the second track, planning to track the same section of the song as in #3 above with another vocal part in that section.
 
What I found in the second track at this point was that there were now 36 take lanes in the second track instead of 9, with clips seemingly randomly spread around the various lanes.  Here is a partial screen shot to give an idea, note the circled areas that show the number of take lanes and highest-numbered take:
 

 
I couldn't fit the entire set of takes in a screen shot due to the number of take lanes, but let me describe some potentially relevant details.  There are exactly 3 clip sections horizontally (i.e. the two in view plus one additional one to the right that is out of view).  All lanes that have clips in them have exactly one clip per lane.  However, there are also 9 lanes (numbers 1-9) that are empty.
 
Note that, to the point where I found this, it was using the original release of Manchester.  I have since updated to the new one, just in case it might prevent future incidents of this type, though there is nothing in the release notes that would seem obviously related.
 
Also note that it did not happen with the first track I'd recorded, only the second one.  There are no recorded tracks below the second one in the project, though there is one more track that does not yet have anything recorded in it.
 
I do not know if the extra lanes appeared before or after doing the post-tracking part in step 3 above.  I'm pretty sure the extra lanes were not there, though, after doing the tracking, splitting, applying trimming, and clip isolation in step 2.
 
Now for the "fun" step of trying to put takes back where they belong and deleting extra lanes so I can start tracking the next part of the song in the second track.  Not fun!!!
 
Anyone else seen this and know a recipe?  I have not seen this in any previous updates, and this style of working is pretty common for me, though I am unlikely to have done it in very recent updates prior to Manchester just due to the timing of updates versus this project.
 

scook
I have not tested it in some time but if you have "Slide Over Old to Make Room" selected in the Track view Options menu > Drag and Drop Options. It could be the source of the problem. Choose one of the other options.

I have Blend Old and New selected there.  There is no horizontal dragging, or dragging anything into areas of track lanes that have anything else in them, though.  In fact, in the track in question, I did no dragging period as that was the first set of takes (all done in succession -- this is partly how I work out the harmony parts), only splitting and deleting of blank areas then applying trimming.  Nor is there any horizontal displacement of the clips, which is, I believe what the feature you've referenced is talking about.  So it just seems like SONAR is thinking all these clips are overlapping and thus needing their own lanes, despite having had trimming applied.  Of course, that still wouldn't explain the set of totally blank lanes.
 
Slight update is that, after updating to the patch release of Manchester, manually consolidating the clips from the messed up track back into 9 lanes, deleting the extra lanes, going to the next portion of the song and doing 12 takes of that section with the part for that section, then doing the same splitting and trimming as with the others, neither of the two tracks I've been working on have added extra lanes.  Of course, I haven't gone back to the first track to record other sections yet (I have to record a third track in the same section first, but it's time to make some dinner now), and I don't have any clue exactly what "caused" (i.e. in terms of user operations, as opposed to underlying behavior) the extra lanes to appear in the first place.

Yeah, as I was doing more parts after dinner tonight, I did find it was early dragging the newer takes down to the earlier lanes and deleting the newer, now empty, lanes.  That did make things a little easier.  But you're right, that I'd prefer things to be inverted as there is, at least for me, a natural order of things where the first few takes are trying to sort out what part I want to do, the next few are refining it and getting more solid, and then at some point there is a point beyond which things start to suffer from getting tired, but maybe there weren't enough of some parts that took longer to figure out, so I still am needing to add a few more parts to get enough choices.  (I usually then comp things down to doubles of each part, at least for background vocals, and sometimes for lead vocals in choruses.)  Having the early takes up at the top helps me more easily see the flow of things because some parts will have quite a few more takes than others, and I'm ultimately getting rid of the clips I don't want as I sort through things, and moving those up to the top (or to another track for the double).
 
The good news is that, in tracking the rest of the song tonight (two more choruses, one with three parts and the other with just two parts, typically 5-6 takes apiece), I had no further incidents with the multiplying clips.  I am now wondering if the Manchester patch 1 cured whatever weirdness caused the earlier problem, though neither of the two fixes mentioned sounds remotely likely to be involved.  Very strange indeed, but at least I'm done tracking this song and will next be on to comping the BGVs then mixing.

Interesting.  Well, I'm hoping it doesn't come back.  I didn't use Kingston at all, but used whatever the "L" one was (my memory isn't working at the moment apparently) and had problems with crashing on using take lanes, but that was with MIDI.  I think that is supposed to be cured in Manchester, but I haven't had time to go back to the sample I set up for a recipe to check it.  I also save a backup copy of my projects fairly frequently.  Pretty much after any major milestone, for example after finishing takes of a section.  I haven't tried going back to my saved versions to see if I can localize where the issue crept in, but it may well have been between milestones anyway.  I ended up manually cleaning things up (albeit with the resultant take lanes in non-numerical order, and it didn't recur in the rounds of additional takes I did later last night.