Separate positive rate mixtures fit 53 of 54 network summaries within a conservative normal-error cutoff

We tested each of 54 published network summaries against any separate positive mixture of 40 adjustment rates and a zero response. A linear program finds the smallest maximum marginal-SE adjustment to the fast, slow and spatial slopes. With spatial-scale error included, 53 networks are compatible at a 3.608 cutoff from a 162-coordinate normal-error rectangle. Three exceed 1.96. The single large mismatch is a benthic network with an unusually large slow response. It requires an 8.163-SE adjustment, and a simple slope inequality independently gives a 7.945-SE lower bound. Dropping the target-gain limit removes that mismatch. Independent polygon tests pass all 196 checks. The two summaries generally do not rule out separate lag mixtures as an explanation for weak tracking. This does not establish a lag mechanism. Published error scales, positive-slope selection, a finite rate grid, source exposure proxies and absent raw trajectories limit inference. The cutoff is a conditional diagnostic, not a calibrated rejection probability.