Having done a few lactate tsts with both runners and swimmers, I have seen lots of variations in the actual blood lactate associated with what I would consider to be threshold. As B2 has pointed out ther is often a trend but no true deflection, and for followup tests I support using a standard blood lactate value to see at what % VO2max, what % HR max and what running speed is associated with a reference lactate (4.0 mmole for example). Whether or not 4.0 is your threshold you can easily watch where the regression curve passes through 4.0 and what speed that is associated with. My feeling is a true threshold is an intensity of effort that will produce a "steady state" of blood lactate accumulation (be it 3.0 or 7.5 mmole -- which is about the range many studies have shown the deflection to occur). A typical lactate profile involves a fairly short time at a progressively faster series of intensities, and look for a deflection, which can be quite subjective. Some people look for 2 different thresholds -- 2.0 mmole and 4.0 mmole. But once you have what you consider to be a deflection then you go back and try a steady 20minute effort at that supposed threshold. One of 3 things can happen -- during this 20min run periodic blood samples will show (1) a gradual increase in lactate accumulation, (2) a steady drop in blood-lactate accumulation , or (3) a relative steady state of blood-lactate accumulation. #3 seems more like a true threshold to me -- a point where clearance keeps up with production. Because of the different ways of testing and interpreting a deflection, etc. I think Luv2run has it best by using 15k race pace (for good runners at least). However, this is not too good for slower runners who take an hour to race 10K, because the pace at which you can race for about 1 hour is a better way of explaining it. Some swim coaches use a T-50 test which is the speed you can hold for an all-out 50min swim. Others use a T-30 test, with a 30-min effort. In well-trainined runners I get 86 to 88% of VO2max over and over again with a range of about 82-91% (which is a whole lot less variation than is what you can see in lactate values associated with a deflection (2.6 to 7.6)) I still like using performances to determine training intensities, because performance is what it is all about, and I am pretty convinced that two runners who both race 13:20 for 5k are pretty equal, even though one may have a threshold associated with 4 mmole and the other 6.5mmole blood lactate. Performance takes into account psychological factors as well as physiological -- so base training on performance.