Skip to contents

This vignette sets out the mathematics behind the models used in mustelus, step by step. It is intended as a reference for the other vignettes, which state results without deriving them.

The von Bertalanffy growth function

The growth vignette introduces the von Bertalanffy growth function as a differential equation in which growth rate declines linearly as length increases:

dLdt=K(L∞−L)\frac{dL}{dt} = K(L_\infty - L)

For an individual animal, age and time advance together, so the rate of change in length with time is the same as its rate of change with age, aa. Written in terms of age, the equation is

dLda=K(L∞−L)\frac{dL}{da} = K(L_\infty - L)

The aim is to find length as a function of age, L(a)L(a).

The general solution

The equation is separable. Dividing both sides by L∞−LL_\infty - L and multiplying by dada puts the terms in LL on one side and the terms in aa on the other:

dLL∞−L=Kda\frac{dL}{L_\infty - L} = K \, da

Integrating both sides,

∫dLL∞−L=∫Kda\int \frac{dL}{L_\infty - L} = \int K \, da

The left-hand side follows from ∫dx/(b−x)=−ln⁡(b−x)\int dx/(b - x) = -\ln(b - x), which holds while length is below the asymptote, L<L∞L < L_\infty:

−ln⁡(L∞−L)=Ka+c-\ln(L_\infty - L) = Ka + c

where cc is a constant of integration. Multiplying both sides by −1-1 and exponentiating,

L∞−L=e−ce−KaL_\infty - L = e^{-c} e^{-Ka}

Writing A=e−cA = e^{-c} for the constant and rearranging gives the general solution:

L(a)=L∞−Ae−KaL(a) = L_\infty - A e^{-Ka}

The distance remaining to the asymptote, L∞−LL_\infty - L, decays exponentially with age at rate KK. The constant AA is fixed by any one known point on the curve, and the choice of that point is what distinguishes the two forms of the equation.

In terms of t0t_0

Beverton and Holt (1957) fixed the curve by the hypothetical age at which length would be zero, t0t_0. Setting L(t0)=0L(t_0) = 0 in the general solution,

0=L∞−Ae−Kt00 = L_\infty - A e^{-K t_0}

A=L∞eKt0A = L_\infty e^{K t_0}

Substituting this back into the general solution,

L(a)=L∞−L∞eKt0e−KaL(a) = L_\infty - L_\infty e^{K t_0} e^{-Ka}

and collecting terms gives the familiar form:

L(a)=L∞(1−e−K(a−t0))L(a) = L_\infty\left(1 - e^{-K(a - t_0)}\right)

In terms of L0L_0

The curve can instead be fixed by length at age zero, L0L_0, which for a chondrichthyan is length at birth. Setting L(0)=L0L(0) = L_0 in the general solution, and noting that e0=1e^{0} = 1,

L0=L∞−AL_0 = L_\infty - A

A=L∞−L0A = L_\infty - L_0

Substituting this back into the general solution,

L(a)=L∞−(L∞−L0)e−KaL(a) = L_\infty - (L_\infty - L_0)e^{-Ka}

Writing L∞L_\infty as L0+(L∞−L0)L_0 + (L_\infty - L_0) and collecting the terms in L∞−L0L_\infty - L_0 gives the form used by growth():

L(a)=L0+(L∞−L0)(1−e−Ka)L(a) = L_0 + (L_\infty - L_0)\left(1 - e^{-Ka}\right)

Converting between the two forms

Both forms are the general solution with a different expression for AA, and they describe the same curve when the two expressions are equal:

L∞eKt0=L∞−L0L_\infty e^{K t_0} = L_\infty - L_0

Solving for L0L_0,

L0=L∞(1−eKt0)L_0 = L_\infty\left(1 - e^{K t_0}\right)

and dividing by L∞L_\infty and taking logarithms to solve for t0t_0,

t0=1Kln⁡(1−L0L∞)t_0 = \frac{1}{K}\ln\left(1 - \frac{L_0}{L_\infty}\right)

Because length at birth lies between zero and the asymptote, the term inside the logarithm lies between zero and one, and t0t_0 is negative. For the example curve in the growth vignette, with L0L_0 = 73 cm, L∞L_\infty = 265 cm and KK = 0.14, this gives t0t_0 = -2.3 years.

Checking the solution

Differentiating the L0L_0 form with respect to age,

dLda=K(L∞−L0)e−Ka\frac{dL}{da} = K(L_\infty - L_0)e^{-Ka}

From the general solution with A=L∞−L0A = L_\infty - L_0, the distance remaining to the asymptote is L∞−L(a)=(L∞−L0)e−KaL_\infty - L(a) = (L_\infty - L_0)e^{-Ka}. Substituting this in recovers the original equation:

dLda=K(L∞−L(a))\frac{dL}{da} = K\left(L_\infty - L(a)\right)

Time to grow half of the remaining distance

The growth vignette notes that an animal takes ln⁡(2)/K\ln(2)/K years to grow half of the remaining distance to L∞L_\infty, whatever its current length. From the general solution, the distance remaining at age aa is L∞−L(a)=Ae−KaL_\infty - L(a) = A e^{-Ka}. After a further interval Δ\Delta it is

L∞−L(a+Δ)=Ae−K(a+Δ)=(L∞−L(a))e−KΔL_\infty - L(a + \Delta) = A e^{-K(a + \Delta)} = \left(L_\infty - L(a)\right)e^{-K\Delta}

so in any interval of length Δ\Delta the remaining distance shrinks by the same factor, e−KΔe^{-K\Delta}. Setting that factor to one half,

e−KΔ=12e^{-K\Delta} = \frac{1}{2}

Δ=ln⁡2K\Delta = \frac{\ln 2}{K}

which depends on neither age nor current length. With KK = 0.14, the remaining distance halves every 5 years.

References

Beverton, R.J.H. and Holt, S.J. (1957) On the Dynamics of Exploited Fish Populations. Fishery Investigations Series II, Volume 19. Ministry of Agriculture, Fisheries and Food, London.

von Bertalanffy, L. (1938) A quantitative theory of organic growth (inquiries on growth laws. II). Human Biology 10, 181–213.