Sequential Transit Chains and Plug-Flow Limits

Why Represent Transit as a Chain?

A single well-mixed compartment gives material an exponential distribution of exit times: some leaves quickly, while a tail remains for a long time. A sequence of transit compartments changes that shape. Material must pass through each stage in turn, so it cannot reach the end immediately; increasing the number of stages concentrates arrival times around their mean. This provides a useful bridge between a broad, well-mixed timing distribution and the idealized fixed delay called plug flow.

Here, the question is specifically about the time for a bolus to traverse a sequence of transit stages. There is no absorption, degradation, or branching in the timing calculation. Those processes can change how much material remains available to arrive, and must be represented separately if they are included.

The standard assumptions for the chain are: each stage is well mixed; transfers are one-way and sequential; every stage has the same constant transit rate; and the initial bolus has no spread in its entry time. These assumptions are about timing, not about the anatomical claim that the gut literally consists of identical compartments. We choose the rate so the chain has a specified mean transit time \(\tau\):

The units of \(k_{\mathrm{tr}}\) are inverse time. Thus a larger \(N\) at fixed \(\tau\) means that each stage has a faster rate, while the total expected time through all stages remains unchanged.

From Compartment Balances to Arrival Times

Let \(Q_i(t)\) be the amount in stage \(i\), and let the initial dose be \(D\), so \(Q_1(0)=D\) and \(Q_i(0)=0\) for \(i>1\). For \(N>1\), the mass balances are

The final outflow rate is \(k_{\mathrm{tr}}Q_N(t)\). With no loss, the cumulative fraction arrived by time \(t\) is one minus the fraction still in the chain. Equivalently, the arrival time \(T\) of a randomly selected unit of the bolus is the sum of the residence times in the \(N\) stages. Under the assumptions above, each stage residence time is independent and exponentially distributed with rate \(k_{\mathrm{tr}}\).

A short derivation makes the role of the stages explicit. For one stage, \(f_1(t)=k_{\mathrm{tr}}e^{-k_{\mathrm{tr}}t}\). If \(f_{N-1}\) is the density for the first \(N-1\) stages, independence means that the density after one more stage is their convolution:

Suppose \(f_{N-1}(u)=k_{\mathrm{tr}}^{N-1}u^{N-2}e^{-k_{\mathrm{tr}}u}/(N-2)!\). Substitution gives

This proves the density by induction from the one-stage case. Integrating the density gives the stated cumulative distribution function; it can also be verified by differentiating that expression and checking that it is zero at \(t=0\). The mean and variance follow by adding the means and variances of \(N\) independent exponential stage times, each with mean \(1/k_{\mathrm{tr}}\) and variance \(1/k_{\mathrm{tr}}^2\).

Worked Arrival-Time Calculations

How the Chain Approaches Plug Flow

In plug flow, every unit has transit time exactly \(\tau\). Its arrival distribution is a point mass at \(\tau\): the cumulative fraction is zero before \(\tau\) and one at or after \(\tau\). A finite chain is not plug flow. It has a continuous distribution of arrival times, with some material arriving earlier and some later than the mean.

The relationship between these representations follows directly from the variance. With \(k_{\mathrm{tr}}=N/\tau\), the chain has mean \(\tau\) for every \(N\) and variance \(\tau^2/N\). For any fixed \(\epsilon>0\), Chebyshev’s inequality gives

Thus, as the number of equal stages grows while the mean is held fixed, the arrival time concentrates around \(\tau\). This is convergence in probability to the fixed plug-flow delay. It does not mean that any finite chain has a hard minimum arrival time of \(\tau\); its distribution still allows earlier arrivals.

Choosing and Checking a Timing Representation

A single well-mixed compartment is the \(N=1\) case and has the broadest timing distribution among these equal-mean chains. A finite chain can represent a progressively narrower distribution while retaining a smooth arrival profile. Plug flow is the limiting idealization when timing dispersion is negligible for the question being modeled. These alternatives are not interchangeable: they make different predictions about early and late arrival, even when their mean transit times agree.

Useful checks follow from the model’s assumptions and equations. First, confirm the rate and units: \(k_{\mathrm{tr}}=N/\tau\) must have units of inverse time. Second, verify that the computed CDF is between zero and one, is nondecreasing, and approaches one as time increases. Third, check that the mean remains \(\tau\) when \(N\) changes and that the standard deviation decreases as \(1/\sqrt{N}\). In a compartment implementation, total chain mass plus cumulative outflow should remain equal to the initial dose, up to numerical error.

A common mistake is to hold \(k_{\mathrm{tr}}\) fixed while increasing \(N\), then claim that the mean transit time is unchanged. It is not: with fixed rate, the mean is \(N/k_{\mathrm{tr}}\), so it increases with the number of stages. To compare shapes at a fixed mean \(\tau\), the rate must instead be set to \(N/\tau\). A related error is to call the finite chain “plug flow” because its mean is fixed; only the limit has zero timing variance.

The chain is limited by its equal-rate, independent-stage assumptions. A real transit process may have unequal segment times, variable entry times, feedback, or other sources of heterogeneity. A narrow chain or plug-flow approximation can then understate the timing spread; a broad one can overstate it. Also, the distribution here describes arrival at the chain’s endpoint in the absence of loss. If absorption occurs during transit, the amount reaching the endpoint and the timing of absorbed drug require a model that includes that process; endpoint arrival alone does not determine systemic appearance.