This is the last of the Food for Thought comments on Imperfection. (So cheer up!)The final sorcue of imperfection to be discussed is dispersion or non-concentration. Simply is the scatter of parts that form a unit of occurrence that leads to a mismatch of contact per part and a spread of non-simultaneous contacts across parts. Other approaches of analysis have been offered, such as search theory, see survey of Pissarides (2000), Diamond (1984), Howitt (1986) and Farmer (2010). They find multiple positions in a continuum of equilibria (indeterminancy of solution). Here,in this analysis, I use queueing theory, see Gross and Harris (1998) and spectral theory, see Kuttler (2010) to present and specify the presentation. The purpose is still to show how reality affects the Keynesian concept of effective equilibrium stasis.The properties of dispersion of parts are rank of queue and phase of spectrum. Parts, as explained in previous comments, are formed as trials of occurrence drive towards completion a unit of occurrence. As the completion process approaches its limit, parts converge to a degree of concentration, one of the sorcues of unit perfection. The convergence is limited by complexity since the forces of completion adjustment get entangled, inducing an entropy of reduction. (No more in this comment!). Parts queue in order to reach a contact with other parts along the path of completion and they have a rank of wait (storage) as they get bumped up the ladder of the queue until served. This is a vertical generation process, reflecting a matching technology. The rank of a part rises in the queue (the queue is shortened for a part) as a prior reaches contact and shifts (interval) with other parts along the path of unit completion. The mismatch of wait, for a part with a higher rank, and the unit dispersion are reduced or concentration increases and parts converge.On the other hand, the spectrum of parts consists of phase intervals (distances) among parts. Each phase interval increases as the queue of parts is shortened. This is an horizontal allocation process, reflecting a spectral transformation function that depends on the phase interval. The phase interval of the spectrum rises and the dispersion increases and parts diverge horizontally. Thus, a rise in the rank queue per part implies a reduction in dispersion and a rise in the phase interval of the spectrum across parts implies an increase in dispersion. In order to assure the stability of the process, we can close the square of nonlinear dialectics of interdependent properties of parts. When the queue is shortened (as the rank of the part rises), the phase interval increases and when the phase interval decreases the queue line is lengthened (and the rank is reduced).At a given level of completion, the unit occurrence has a mix of rank of queue per part and phase of spectrum across parts which corresponds to a degree of dispersion. We can specify the above, with a simple model of dispersion based a) on a matching technology per part as the square root of the wait period subject to a Poisson generation process of the number of contacts in the wait period. Furthermore, the model is based on b) a spectral transformation of the phase interval as the squared interval (distance) across parts subject to an exponential distribution of the intervals. Under this sorcue of imperfection, unit operation and behavior and its variation feedback is bounded by mismatches between parties in market trades and non-simultaneous transactions that require the holding of liquid assets. Furthermore, there are multiple positions of equilibria rather than a unique equilibrium that clears markets. Spreads of completed transactions are persistent. This describes another form of a Keynesian concept of equilibrium that corresponds to an effective stasis that could differ from the Pareto optimum general equilibrium position.The sorcues of imperfection can be combined in a model of the mode of operation, behavior and feedback that is consistent with the effective equilibrium concept of Keynes.