There was no question about it when I looked in the dye vats after the first round.  I had pink.  Not bright Barbie-doll-pink (thankfully).  A sort of greyish-plaster-pink in the pale vat ... maybe more like dull purple in the other two.  Unfortunately pink happens to be one of my least favourite colours.  I had about six metres of cloth divided between three vats and it had taken two or three days to stitch and bind all the resists ... definitely time for a major rethink.  

First step - figure out how a dye mix that was meant to result in a greenish brown had ended up pink.  If anything the mix had looked too green when I made up the vats.  The two purplish vats probably reflected how I'd adjusted the mix accordingly.  My guess was that the lemon yellow dye had been shamefully bullied by the (tiny amount of) magenta and wimped out on me.  Which reminds me (too late) why I usually prefer golden yellow.  

Yes, dyes have their own personalities.  Red (especially magenta) is a bully, reacts quickly and easily overwhelms the others.  Blue is slow to react, needing twice as long to reach a full depth of colour, but quite resilient once it's got there (try removing an unwanted turquoise stain).  Yellow (especially lemon) reacts quite quickly but is a wimp and is easily overwhelmed by the others.  

It's relatively easy to compensate for all this when you dye with pure colours, but there are good reasons for using a mixed colour in shibori.  I've worked out most of my mixes using thickened dyes (more convenient) but I'm finding that a mix that works fine that way can behave differently in the vat.  I think this is because of differences in the way the dye reacts with the cloth in each case.  The daunting implication is that I am going to have to figure out a system to re-test a lot of my mixes in the vat.  

Meanwhile, I needed to figure out how to deal with all this pink.  My original plan had been three blue-black vats, which should have given me a range of dark blues, browns and greenish neutrals.  But to cancel out the pink tone, I needed some green.  I decided to substitute turquoise for the royal blue I had planned to use ... and crossed my fingers.  

On the whole it worked.  I didn't of course get the colours I'd planned ... I ended up with a lot more blues ... and I still have more pinkish tones than I want so may be over-dyeing some of these.  Overall not the most successful results - although, as ever, there are some stand-out pieces.  

I was also using this dye session to experiment with clamped resists, in particular, partially clamping the pleated and stitched pieces.  The results were interesting but on the whole the edge between the clamped and the unclamped section is a little obvious so this needs more thought.  It didn't quite work for what I have in mind.  The other experiment was to try out ways of restricting the pleated shibori to just one section of cloth - without resorting to wax or similar.  

The colours were really not good, so I've greyed them out to focus on the marks.  The one on the right is quite exciting with lots of potential for development.  Guess what, this was the only sample in the whole batch with lemon yellow on it ... in all the wrong places!