Marginal Value Theorem

The marginal value theorem (MVT) is a foundational mathematical optimality model in behavioral ecology and Optimal Foraging Theory (OFT) formulated by Eric L. Charnov in 1976 (and applied simultaneously to male mate-guarding and copulation duration in yellow dung flies by Geoffrey A. Parker and R. A. Stuart in 1976). Whenever an animal, root system, or human hunter-gatherer forages in a patchy environment—such as a bumblebee visiting flower inflorescences, a great tit searching pine cones for caterpillars, a diving seal hunting fish at depth, or a parasitoid wasp ovipositing in clusters of host larvae—its own consumption (or oviposition) progressively depletes the local patch, causing its instantaneous rate of resource capture to decelerate with time spent in the patch (t). How long should the forager remain in the current depleting patch before paying the energetic and time cost (travel time τ) of moving to a fresh, undepleted patch? Charnov proved that to maximize its long-term net rate of energy intake (E_net) across an environment of many patches, an optimal forager must depart from its current patch at the precise moment (t^*) when its instantaneous marginal capture rate inside the patch drops to the long-term average rate of net energy gain across the entire habitat (including inter-patch travel time τ) (Charnov, 1976; Parker & Stuart, 1976; Stephens & Krebs, 1986).

Consider a forager moving sequentially through a habitat containing k discrete patch types (with encounter proportions p₁, p₂, …, p_k, where ∑ p_i = 1), separated by an average inter-patch travel time τ during which the animal finds no food and expends metabolic power e_T. Once inside a patch of type i, the forager's cumulative net energy assimilated after residence time t_i is given by the gain function g_i(t_i) (gross energy captured minus metabolic foraging cost inside the patch).

Because the forager depletes easily accessible prey first (or because prey hide upon detecting the predator), the cumulative gain curve g_i(t_i) is monotonically increasing (g_i′(t_i) > 0) and concave downward (g_i″(t_i) < 0)—meaning the marginal capture rate ∂g_i / ∂t_i diminishes continuously with patch residence time t_i.

By the renewal-reward theorem, the long-term average net rate of energy intake (E_n) over many patch visits is the ratio of expected net energy gained per patch cycle to expected total time per cycle (travel time + patch residence time):

E_n = [ ∑ p_i g_i(t_i) − e_T τ ] / [ τ + ∑ p_i t_i ]

To find the optimal patch residence time t_j^* that maximizes E_n, we set the partial derivative ∂E_n / ∂t_j = 0 using the quotient rule:

∂E_n / ∂t_j = p_j [ (∂g_j / ∂t_j)(τ + ∑ p_i t_i) − (∑ p_i g_i(t_i) − e_T τ) ] / [ τ + ∑ p_i t_i ]² = 0

Rearranging the bracketed numerator immediately yields Charnov's Marginal Value Theorem:

( ∂g_j / ∂t_j ) |_{tj = tj^*} = E_n^* = [ ∑ p_i g_i(t_i^*) − e_T τ ] / [ τ + ∑ p_i t_i^* ]

In a homogeneous environment where all patches have the same gain curve g(t) (and setting travel cost e_T = 0 by measuring net energy relative to travel metabolism), this optimality condition reduces to the famous Charnov tangent equation:

g′(t^*) = g(t^*) / ( τ + t^* )

Graphical Tangent Construction and Four Core Predictions

Plotted with travel time τ extending to the left of the origin (−τ on the horizontal axis) and patch residence time t extending to the right of the origin, the quantity g(t) / (τ + t) is the slope of a straight line drawn from the point (−τ, 0) on the negative time axis to the point (t, g(t)) on the gain curve. That slope is maximized when the straight line is tangent to the gain curve g(t), yielding four quantitative predictions (Charnov, 1976; Stephens & Krebs, 1986; Brown, 1988):

1. The Travel-Time Effect (τ ↑ ⇒ t^* ↑): As the distance or travel time τ between patches increases, the origin of the tangent line (−τ, 0) shifts farther to the left, flattening the tangent slope E_n^* and shifting the point of tangency (t^*) farther up the curve to the right. Thus, when travel between patches is long or difficult, a forager should stay longer in each patch and deplete it more thoroughly before leaving.
2. Equalization of Marginal Giving-Up Rates Across Heterogeneous Patch Types: In a habitat containing rich, intermediate, and poor patch types (g₁(t), g₂(t), g₃(t)), the forager should stay longer in richer patches inside the same habitat, but must leave every visited patch type at the exact same marginal capture rate g_i′(t_i^*) = E_n^*—reducing every patch in the habitat down to the same Giving-Up Density (GUD) of remaining prey (Brown, 1988).
3. Habitat-Wide Enrichment Shortens Residence Time: If the overall habitat becomes richer (E_n^* increases because travel time τ decreases or neighboring patches become more productive), the tangent line steepens, and the forager should leave a focal patch of fixed quality sooner (t^* decreases).
4. Exclusion of Poor Patches: Any patch type j whose initial maximum capture rate at t_j = 0 is lower than the habitat's long-term average gain rate (g_j′(0) ≤ E_n^*) should be ignored and bypassed altogether.

The marginal value theorem has been tested across hundreds of laboratory and field studies spanning foraging, reproduction, diving physiology, and cognitive neuroscience:

1. Charnov, E. L. (1976). Optimal foraging, the marginal value theorem. Theoretical Population Biology, 9(2), 129–136. https://doi.org/10.1016/0040-5809(76)90040-X90040-X)
2. Parker, G. A., & Stuart, R. A. (1976). Animal behavior as a strategy optimizer: Evolution of resource assessment strategies and optimal emigration thresholds. The American Naturalist, 110(976), 1055–1076. https://doi.org/10.1086/283126
3. Cowie, R. J. (1977). Optimal foraging in great tits (Parus major). Nature, 268(5616), 137–139. https://doi.org/10.1038/268137a0
4. Kacelnik, A. (1984). Central place foraging in starlings (Sturnus vulgaris). I. Patch residence time. Journal of Animal Ecology, 53(1), 283–299. https://doi.org/10.2307/4357
5. Stephens, D. W., & Krebs, J. R. (1986). Foraging Theory. Princeton University Press. https://press.princeton.edu/books/paperback/9780691084428/foraging-theory
6. Brown, J. S. (1988). Patch use as an indicator of habitat preference, predation risk, and competition. Behavioral Ecology and Sociobiology, 22(1), 37–47. https://doi.org/10.1007/BF00395696
7. Hayden, B. Y., Pearson, J. M., & Platt, M. L. (2011). Neuronal basis of sequential foraging decisions in a patchy environment. Nature Neuroscience, 14(7), 933–939. https://doi.org/10.1038/nn.2856