Graphical representation of a Lorentzian spectral density, eqn (1.7), as the Fourier transform of monoexponential autocorrelation functions, eqn (1.6), for different values of the correlation time τ_c. The crossover from the plateau (ω_kτ_c≪1)≈2τ_c at low angular frequencies to the limit (ω_kτ_c≫1)≈2/(ω$k2$τ_c) at high angular frequencies occurs around the positions ω_k=τ$c−1$. Note that for ω_k<τ$c−1$, the spectral density (ω_k) increases with increasing values of τ_c and decreases in the opposite case ω_k>τ$c−1$. This is exemplified by the vertical lines and the dots at two angular frequencies complying with the respective conditions ω_a<τ$c−1$ and ω_b>τ$c−1$ in the frame of consideration here. Qualitatively, this behaviour applies generally to all stochastic processes irrespective of the actual shape of the autocorrelation function. With respect to field-cycling NMR relaxometry, this means that spin–lattice relaxation rates 1/T₁ increase with longer τ_c values (i.e. slower fluctuations) for ω_kτ_c<1 whereas they decrease for ω_kτ_c>1.