Having never attempted the dropper-method (I'm sceptical), and to possibly prevent too much time-investment for little benefit, may I suggest a preliminary "trial-run" to test the methodology?
Although water bottles may work fine (maybe not), it might work better to use rinsed plastic milk jugs with uniform volumes of water.
I realize plastic milk jugs are opaque. So, if they're not sufficiently clear to directly view/compare coloration variations, maybe you can run your dilution-ratios in the milk jugs (allowing greater dilutions and error-margins) and then transfer a portion of dyed water from each jug into clear glass cups for side-by-side comparisons (returning the samples back to the respective jugs between comparison-intervals).
If you prefer water-bottles for initial dilutions, go ahead. But, the typical "labeled" field-use ratio for liquid-dyes is 1:1,303,404 (= 1qt/ac-ft). If 132 drops = 1 oz (using my calibrated dropper, with each dropper being different), then 1 drop would treat 77.14 gal. at "labeled rates". Using 10x the standard (though subjective) labeled-rate would require adding 1 drop to 7.7 gallons of water. 100x "labeled-rate" = 1 drop to 9.86 oz.
Just trying to put things into perspective before you tackle this task. I'm simply guessing that color-variances will be more easily discerned at higher dilution-ratios; thus a greater volume of water per drop of dye.
It might be worth a trial-run with a couple of dyes that you may already have on-hand before casting the net for numerous samples.
