Curso, that's one of these things I've been pondering. If you can get that different brix reading from two various leaves from one plant where's the middle ground? You go for the lowest, the highest or you settle for between 13 and 22?
Brix readings are funny things, that's for sure. In Gray's case he takes readings form the same location on the plant for the most part. He selects "similar leaves" for each and every reading. I haven't pulled out the refractometer in a while on my latest grow. Maybe I will later today.
If you want Brix readings to be a fair and accurate reading of how the plant is doing you should also take samples at the same time in the watering cycle as well as from a similar location. A recently watered plant has a lower brix than one that is on the verge of wilting from need for water. You also need to have a consistent method of taking the sample. If you do all those things brix readings give you a reasonable indicator of how the plant is hooked up in the soil.
Higher brix is an indicator that the plant's metabolism is firing on all cylinders - the plant is efficiently mobilizing all the mineral ions that it needs from the soil to make the most carbon energy it possibly can via photosynthesis - lots of sugars made by the sun, and lots of mineral cofactors available in the plant's sap.
Bad test results from bad sampling practices are just that - bad results. In Gray's case he replicates the test to the best of his abilities each time he takes a reading and his data is a cut above most of the other brix data I've seen posted.