I'd say that on the fd curve for the 40# @ 30" yew, neither the straight line nor the curve are anywhere near a fit for the data.
Allowing for measurement error It's pretty straight. I'd tend to discount to zero draw point as pretty irrelevant and easilly subject to error*.
The blue line is my take on it....
The slight initial deviation from the expected could be due to a variety of things, miss-measurement, string stretch, poor scale accuracy at low weight. Theoretically the curve will pass through the brace height at zero pounds... but it's close enough for jazz.
The slight upward deviation as it aproaches full draw could easilly be the string angle change.
After all there is nothing in the geometry of a longbow that would make you expect a linear force draw curve, the constantly changing string angle. The large deflection of the limbs where most physics confines it to 'small angles of deflection'. The material taken to close to the elastic limit (and in some csaes beyond.
It's a minor miracle that it's anywhere near linear.
Del
*Does the tiller rig have the scale hanging on the sting? If so, then at brace height the scale should be reading the weight of the scale and it will be giving a small deflection. My scale weighs about 2# so I "zero" the scale at 2#