Dependency values are used to determine if the fitting equation is over parameterized or if the selected fitting function does not describe the data very well. A dependency of 1 indicates that the parameter value may not be very reliable. The dependency value is meaningful once the chi^2 value has been reduced. For example, open \origin40\samples\fixexmp1.opj. This is 1st Order Exponential Decay data. Enter the Non-Linear Curve Fitter and select the ExpDecay1 fitting function. Now, select Action:Fit and allow all fitting parameters to vary. After fitting you will see that x0 and A1 have a dependency of 1. This is because the function can be rewritten as the following:

y = y0 + A1 * e^( x0/t1 ) * e^( -x/t1)

As you can see, the A1 * e^( x0/t1 ) term can have different A1 and x0 pairs of values to produce the same result. This will cause the two parameters to have a mutual dependency, as indicated by a value of 1. If you now fix (uncheck the vary? box) either A1 or x0, you will see that the dependency is no longer 1.