arXiv:1905.02474v1 [physics.bio-ph] 7 May 2019 a st altealmtisfnaetllw fecology, of as laws go fundamental allometries [1] the Colyvan call & increasing to Ginzburg as attracted far years. has recent mass in body attention animal with rically population use- in the relations demonstrate scaling modeling. mass conclusively By allometric to cycles. of of wish population fulness implications animal we of the so, period at doing the look for also size and will body mass We the Prey of quantities: mass. parameters two all predator only our rewrite of is to terms able It in equations be populations. will we actual used that of be hope behaviour can model the idealised predict highly to this that way a equations such predator-prey in parame- Lotka-Volterra the classic estimate the to allo- of possible using ters by is useful it ap- that, scaling, argue more makes mass will a we metric that paper, be this model might In from simplified predictions approach. principles a clear general that but it more proximate believe but We derive case, isolated to it. the difficult for be interesting study will be to course wishes pa- of one of system will measurements individual precise every requires for rameters that parame- of model number com- A large due too ters. often predictions the being definitive about up any uncertainties end to make unfortu- models actually is the to result of plicated real- The many many that as possible. nately for to as account is complications to approach world parameters that common and model a terms a estimating ecosystems, develop add real to of desire describe difficulty we can the If is parameters. problem model recurring a ties, ∗ † [email protected] [email protected] h atta ayeooia aibe cl allomet- scale variables ecological many that fact The communi- ecological of dynamics the modeling When ewrs loer,pplto yais rdtr prey predator, dynamics, population allometry, Keywords: 87.10.Ca 87.23.Cc, allomet prediction numbers: of PACS testable usage make wider and a dynamics p for compa population way animal to in the out of models paves turn This length results th the ecosystems. numerical of and about estimate analytical predictions to Our attempt make mass. we we and para scaling, to ecosystems, allometric useful be this h may examine r mass Using we broad animal paper, with a this quantities for In biological parameters theory. estimate from reliably predictions clearer to way A testable. il orIsiue nvriyo oehgn Blegdamsv Copenhagen, of University Institute, Bohr Niels npplto yais ahmtclmdl fe contain often models mathematical dynamics, population In .INTRODUCTION I. pligalmti cln opeao-rysystems predator-prey to scaling allometric Applying il orInstitute Bohr Niels Dtd a ,2019) 8, May (Dated: nra Eilersen Andreas i Sneppen Kim g aso h nml miial,tesaigrelation be scaling to the aver- Empirically, the found animal. with is the scale of to in mass tends peaks age animals between regu- such elapsed of a time abundance The exhibit - cycle. lemmings population most lar - as animals such many rodents particular, prominently In scaling. allometric of [3]. work al. the foun- see et theoretical West laws, a scaling of quarter-power finding these at for mass. attempt dation the body write compelling animal to a of exploit For terms will we in respectively. that equations Lotka-Volterra 3/4, relationships example, and these For mass is 1/4 with It (approximately) paper. correlate of this rate powers in metabolic to of and use time make generation scaling, [2] allometric will Peters exhibiting we variables physics. which of in list laws a Kepler’s compiled to them comparing ht o eti rdtrpe asrto,i ilhv a have period will a with it find cycle ratios, and limit mass system predator-prey dynamical certain the for of that, anal- stability they linear place, deter- the in measurable to model yse from the law possible With scaling be quantities. this biological should of it coefficients mass that the predator mine and to mass power, prey they some of interaction Also, to ratio predator-prey the the response. with of scale functional should strength III the or that II argue Their predator type the Holling giving a population, satu- prey. prey will increasing and reproduction with predator rate predator the that for plant assumes (whether model substituted resource Lotka-Volterra animal) and generalised or consumer to with similar equations, equations of tem relationship. this of basis theoretical osmr(rdtr asand mass (predator) consumer oe ilb oal ipie n u praht the to approach our and simplified notably be will model mass. Lotka-Volterra , ntelre csse cl,teeaeas examples also are there scale, ecosystem larger the On ozs&Ins[]uetems oprmtrs sys- a parameterise to mass the use [5] Innes & Yodzis ewl eepoeddw iia ah huhour though path, similar a down proceed here will We † ∗ wtealmti cln fanme of number a of scaling allometric the ow eeiepplto yaia models. dynamical population meterise j1,20 øehv ,Denmark. Ø, København 2100 17, ej s. neo ytm ol epu obtain us help would systems of ange eraoal odt rmanumber a from data to reasonably re ai fpe opeaosi real in predators to prey of ratio e i cln osmlf mathematical simplify to scaling ric o ayprmtr ob easily be to parameters many too plto ylsa ucinof function a as cycles opulation T ∝ m 0 . 26 T ∝ 4.W iht ru o a for argue to wish We [4]. m m 1 C / R 8 m stersuc (prey) resource the is 1 R / 8 where , m C sthe is 2 predator-prey functional response will be different. The [7]. Here, x denotes prey, y predator, α is the per- original Lotka-Volterra equations on which we will be capita reproduction rate of prey, and δ is the per capita basing our model assume that the predation and preda- death rate of predators in the absence of prey. The in- tor reproduction rates increase in proportion with prey teraction strengths β and γ are slightly harder to define. population density, a so-called Holling type I functional β denotes the risk of each prey being eaten per predator, response. We here assume that prey population is always and γ represents the increase in predator reproduction far from the carrying capacity of the ecosystem, resulting rate per prey. These latter two parameters are of course in a prey reproduction rate that is also proportional to more difficult to estimate than the first two, and we will prey population. Instead of trying to determine a biolog- therefore need to find a way around this obstacle. ically reasonable coefficient for the scaling of interaction As opposed to Yodzis & Innes, we choose to work with strength with the predator-prey mass ratio, we will let animal abundances rather than biomass densities. We do the coefficient remain unknown. We will determine the this because it is conceptually easier and data are more equilibrium populations in terms of this unknown coeffi- readily available for abundances than for biomass densi- cient. Luckily, it turns out that when we look at the ratio ties for the systems that we wish to study. A complica- of the populations, this coefficient cancels out. Thus, our tion arising from this is that when working with abun- method still yields useful information. dances, there is a distinction between somatic growth (in- Finally, we will look at the period of population cy- dividuals growing larger) and reproductive growth, which cles. The simplest version of the Lotka-Volterra equa- would be unimportant if we were to work with biomass tions has a non-trivial equilibrium which is a center, densities. We shall therefore ignore the finer details of an- rather than a limit cycle. Here, we likewise find a pe- imal reproduction and growth, and simply assume that riod of T m1/8m1/8 as mentioned above. In order to all growth results in the production of new individu- ∝ C R obtain the empirically determined m1/4 relationship with als. Furthermore, we assume that the populations are population cycle length, Yodzis & Innes point out that large enough and reproductive events evenly distributed one can assume a direct proportionality between preda- enough in time that population growth can be modelled tor (consumer) size and prey (resource) size. While this as continuous rather than discrete. relationship may hold in many systems (see e.g. [6]), it According to Peters [2] we then have the following em- certainly does not in such cases as the wolf-moose sys- pirical relation for reproduction rate: tem studied by Peterson et al. [4], and the relationship is hardly well-defined in systems where the resource is a 1 1/4 1 plant. Ginzburg & Colyvan [1] even present a critique α = mx− [day− ] (3) 400 of the whole idea of using only the linearisation of the Lotka-Volterra equations to predict the length of popu- In the cited mass scaling relations, all masses are in lation cycles. It would therefore be preferable if we could kilograms. As the predator-prey pairs we will be examin- derive a relationship between prey mass and cycle period ing here are all mammals, we shall be using the mass scal- that is independent of predator mass. This is what we ing relations that apply to mammals. For cold-blooded will attempt to do in the following. animals such as reptiles the relations will be different, The model put forward here is thus an application though not radically so. of the basic idea of Yodzis & Innes to a heavily sim- It should be possible to calculate the death rate of plified system of equations, without making attempts predators in the absence of prey from the so-called at determining the exact interaction strength between turnover time. This is defined as the average time it will predator and prey directly. It is our hypothesis that even take an animal to metabolise its entire energy reserves. such a simplified model will still give reasonable order- In turn, this can be calculated from the metabolic effect. of-magnitude predictions about real ecosystems. Again from [2]

1/4 II. PARAMETERISING THE tturnover = 19my [day] (4) LOTKA-VOLTERRA EQUATIONS This implies The original Lotka-Volterra predator-prey equations read as follows: 1 1 1/4 1 δ = t− = m− [day− ] (5) turnover 19 y dx The coupling coefficient β we assume to be propor- = αx βxy (1) dt − tional to predator ingestion rate. We believe this to be justified, since the more a given predator consumes, the higher the per-capita risk of being eaten by it should be dy for the prey. The predator ingestion rate in terms of = γxy δy (2) dt − energy scales with mass as 3

will discuss below suggests that the wolves have a much lower efficiency than we would expect based on the above 3/4 1 I m [J (day predator)− ] (6) (η 2 %) [10]. In laboratory experiments, a figure of ∝ y · · about≈ η = 10 % is observed [10], and for lack of a bet- [2]. The number of individual prey that a predator ter estimate, we shall use this so-called ten percent law needs to eat to satisfy this energetic demand is inversely in our calculations. Given that we are not going for an proportional to prey mass, and we therefore write β as exact description of any one particular interaction, we believe that it is justified to use this rough estimate. 3/4 The relation between mass of consumed prey (mx,c) my 1 β = k [(day predator)− ] (7) and mass of produced predator (my,p) is now · mx ·

where k is an unknown proportionality constant. my,p = η mx,c (8) Knowing the equilibrium population of prey or predator · should make it possible to determine k if this is desired. assuming that prey and predator have similar energy Our parameterisation thus deviates notably from content per unit mass. Rewriting this in terms of num- that of Yodzis & Innes, since they assume that the bers of individual predators produced (Ny,p) and prey predator death rate and the interaction strength scale consumed (Nx,c) we get with the ratio of prey mass to predator mass to the power of 3/4 (here converted to abundance rather mx,c mxNx,c Ny,p = my,p/my = η = η (9) than biomass, as was originally used). Strictly speak- my my ing, the ingestion rate of y predators reflects some kind of average prey consumption rate at average prey In the Lotka-Volterra equations, the number of preda- abundance. What we really need here is the slope tors produced per unit time is given by the term of predator kill rate as a function of prey abun- dance. Furthermore, the units of the ingestion rate is Ny,p = γxy (10) 1 1 [J (predator day)− ] [prey (predator day)− ] and · · 1∝ · · not [(predator day)− ] as we need it to be for our units and the number of prey consumed by the term to match. Despite· all this, we still believe that the al- lometric scaling of the ingestion rate is a reasonable ap- Nx,c = βxy (11) proximate measure of the predator’s ability to consume and therefore of the dependence of consumption rate on Thus, we get the following relation between β and γ : prey abundance. We now only need to find a way around not knowing the exact proportionality. mx 1/4 1 γ = η β = k my− η [(day prey)− ] (12) The slope of predator kill rate with prey abundance my · · · · that really constitutes β depends on a number of factors (temperature, prey population density, predator satia- We have now written all the parameters of the equa- tion etc. [2]), and it is probably not possible to make tions in terms of the animal body masses alone, with k a universal estimate of it. Instead, we let k embody all from eq. (7) being the only parameter that remains to these complications and tune it to fit the systems that be determined. However, we can get around this by fo- we will study. As mentioned above, it fortunately can- cusing our attention on the equilibrium predator-to-prey cels out in the final calculation of the prey to predator population ratio instead of the absolute populations. population ratio anyway. The Lotka-Volterra equations have the non-trivial The relation between the number of prey eaten and equilibrium the number of predators produced can be derived ap- proximately if we know the ecological efficiency η of the δ α (x, y)= , (13) predator-prey interaction. The ecological efficiency here γ β  refers to the percentage of prey biomass that is converted into predator biomass. Ecological efficiencies vary con- which is neutrally stable. The equilibrium ratio be- siderably depending on the nature of the interaction [8], tween prey and predator populations is therefore and it is therefore difficult to find an estimate that is both precise and general. For systems with a low preda- 3/4 βδ 21 my tor to prey mass ratio and positive correlation of biomass x/y = = (14) αγ η m  density with body mass, ecological efficiency should be x high (η 35 %) according to a review by Trebilco et This number depends only on the masses. We see that al.[9], which,≈ however, deals with aquatic ecosystems. due to the factor 1/γ this ratio is inversely proportional Lindeman’s original paper similarly shows an efficiency to ecological efficiency, so that if our estimated 10 % ef- that rises with trophic level [8]. On the other hand, a ficiency is a factor 2 too great, we will estimate a ratio case study of the Isle Royale wolf-moose system that we that is half the ”correct” value. 4

III. THE PERIOD OF POPULATION CYCLES 103 Trajectory Mean populations The Jacobian matrix of the Lotka-Volterra equa- LV Equilibrium tions at the non-trivial steady state has the eigenval- Timestep ues (i√αδ, i√αδ), meaning that for small perturbations 102

− ] away from equilibrium, the system will oscillate over time -2 with a period of T = 2π . This leads to the afore- √αδ

mentioned scaling of population cycle period with mass Hare [km 1/8 1/8 1/4 1 T mx my , contrary to the observed T mx . A 10 problem∝ with using linearisation in this case∝ is that the period thus obtained only applies when oscillations are relatively small. Population cycles in actual predator- prey pairs, such as the vole-weasel pair in northern Scan- 100 10-10 10-8 10-6 10-4 10-2 100 102 dinavia, can involve fluctuations over two orders of mag- Lynx [km -2] nitude [11]. When solving the equations numerically, we see that much of the time, the population of prey will be FIG. 1. (Colour online) An illustration of the dynamics of the in a state of slow, exponential recovery, while the preda- predator-prey system. This numerical solution is based on pa- tor population slowly approaches zero. When the prey rameters appropriate for the lynx-hare system discussed be- population recovers, the predator population quickly ex- low. The line shows the trajectory of the system in predator- plodes, initiating a swift collapse of the prey population. prey space, and the circles are all spaced evenly in time, at The collapse phase observed in real rodent cycles does in- a separation of 50 days. The distinction between a fast and deed appear to be notably shorter than the growth and a slow segment of the trajectory can be clearly seen from the peak phases, and the corresponding predator cycles are spacing of the circles. Note also that equilibrium abundances are practically identical to mean abundances, meaning that similarly observed to be very sharply peaked [11, 12]. we can use the two interchangeably. We therefore believe that the dynamics can be realisti- cally modelled as consisting of a slow exponential growth phase and a fast collapse phase. Using this two-timescale ditions, but which does not match observations, as the assumption, we will try to derive an expression for the assumption that initial conditions are close to the equi- period T of population cycles. Splitting more complex librium breaks down in the real systems studied here. In- predator-prey models into slow and fast phases has pre- stead, we assume that the initial conditions are far from viously been done by Rinaldi & Muratori [13]. In the the center equilibrium. For this asymptotic approxima- following, we shall use a similar basic idea, but a differ- tion, the period will depend on initial conditions and the ent mathematical approach and solve for the period T , calculated period matches observations better. rather than maximal abundance as they did. An illus- Starting from a population xmin, the prey population tration of the of the cycle and its fast and slow segments should grow as follows: can be seen in fig. (1). For our derivation, we will use the maximum and minimum prey density of a cycle, which αt should be easily obtainable from observations and avail- x(t)= xmine (15) able in the literature. The slow approach to and subsequent drifting away When predator population is low, prey population from the saddle point at (0, 0) is what takes up the ma- grows unobstructed. After one period of length T , we jority of the orbital period of the system. For this reason, should have the maximal population density we will here attempt to derive an approximate relation for the cycle length by looking at the behaviour around the αT δ α xmax = x(T )= xmine (16) saddle point at (0, 0) instead of the center at γ , β . Although the period of the cycle is mainly determined  The time it will take the population to recover to a by the hyperbolic approach to the saddle point, the os- density of xmax now becomes cillation still happens around the center equilibrium at δ , α . As can be seen in fig. 1, the time average pop- γ β 1 xmax xmax 1/4   T = ln = 400 ln mx [day] (17) ulations are very close to the equilibrium populations at α  xmin  ·  xmin · the center. We therefore do not believe that there is a contradiction between using the saddle point linearisa- 1/4 We thus get the mx -relation found empirically. The tion to determine the oscillation period, but determining above expression should be valid when the amplitude population ratios based on center equilibrium values. of oscillations is very large, so that the period of the Using the linearisation around the center equilibrium, predator-prey cycle is dominated by the slow growth we obtain a period that is independent of initial con- phase which in the Lotka-Volterra model occurs at low 5

] 200 ] 15000

-2 Hare -2 Vole Lynx Weasel 10000 100 5000

population [km 0 population [km 0 0 2000 4000 6000 8000 10000 0 2000 4000 6000 8000 10000 time [days] time [days]

1010 1020 Hare Vole Lynx Weasel

100 100 population (log) population (log) 10-10 10-20 0 2000 4000 6000 8000 10000 0 2000 4000 6000 8000 10000 time time (a) (b)

−2 FIG. 2. (Colour online) (a) A numerical simulation of the Lotka-Volterra equations for lynx and hare. xmax ≈ 180 km ,x0 = −2 −2 xmin ≈ 8km , y0 = 0.3 and k = 1.05 · 10 . The period is just under 2000 days, or 5.5 years, and the average hare density is 51 km−2. Average lynx density is 0.059 km−2. The ratio of the averages is 860 hares per lynx. As can be seen from the logarithmic plot, the predicted predator oscillations are too violent, with extinction of lynx at the cycle minimum. When this does not actually happen, it may be due to the fact that lynx can survive partially on other prey when hare population is low 4 −2 2 −2 [14]. (b) The solution obtained using the masses of voles and weasels. xmax ≈ 10 km , x0 = xmin ≈ 10 km , y0 = 20, and k = 2 · 10−4. We still see a cycle somewhat shorter than the observed, with an estimated T ≈ 2.3 years. Again, the predator oscillation is unrealistically violent. Average vole density is 2100 km−2 and weasel density is 4.6 km−2, giving 460 voles/weasel. predator abundances. Note, however, that we at no point (lepus americanus). Although there has been some doubt have assumed that the population crash should be due as to whether the hare population cycle is driven primar- to the influence of a predator. We just assumed that ily by predation or other factors, there seems to be evi- the crash was fast and did not extend the period length dence that changes in hare mortality are mainly due to or influence the exponential growth phase significantly. predation [23]. The population density of hares oscillates The derivation here should therefore be equally valid if a from around 8 to just under 200 per square kilometer over population crash is caused by e.g. a shortage of food or the 8-10 years long cycle [24]. The average density of lynx 2 an epidemic. Given that in the case of many rodents it ranges from 0.03 to 0.3 km− [25]. is unclear if it is actually predation that drives the cycle To see how well our model fits with observations, we [15], this is a significant advantage. plug the average masses of lynx - on average roughly 11 Another interesting feature of this expression is the kg [20] and hares - roughly 1.6 kg [26] into the equations logarithmic scaling with population maximum-minimum and solve them numerically. We choose initial conditions ratio. Hanski has already hinted at such a scaling rela- corresponding to the density per square kilometer when tion for the vole-weasel system [16]. In his 1991 paper, he hare abundance is lowest (x x = 8 and y =0.3 - 0 ≈ min 0 shows that ln xmax correlates with latitude, and that due to the phase difference between lynx and hare popu- xmin lation oscillations, we let lynx population start out high oscillation period also correlates with latitude. Oscilla- and hare population start out low). We then tune the pa- tion period thus also correlates with the logarithm of the rameter k to obtain the correct ratio between cycle highs maximum-minimum ratio. It is possible that we have and lows. Initial lynx abundance is taken to be slightly found a theoretical explanation for this correlation. above minimum, as the predator cycle will lag behind In the next section, we will demonstrate that eq. (17) the prey. The result can be seen in fig. 2 (a). Our roughly fits the pattern seen in oscillating populations in simulation predicts an average prey to average predator nature, although there is a significant deviation between population ratio that is quite close to the observed val- predicted and observed numbers. For the prey-predator ues. The period is off by about a third, which, given ratios on the other hand, the parameters derived above the simplifications of the model is not a bad estimate. mostly give realistic results. The fact that the population collapse takes such a short time in our simulation contributes to our underestimat- ing the period. In reality, the collapse takes about 1-2 IV. COMPARING THEORY WITH DATA years [24]. The spiky appearance of the graph is also not very naturalistic. However, taking increasing predation The classic example of a system described well by the from other predators, increasing susceptibility to disease Lotka-Volterra equations is the interaction between the and other complicating factors that increase with popu- canadian lynx (lynx canadensis) and the snowshoe hare lation density into account would most likely lead to a 6

System my [kg] mx xmax xmin Observed x/y ratio Theoretical x/y Obs. T [days] Theoretical T Lynx-hare 11 ± 1 1.6 ± 0.1 180 ± 80 8 ± 4 600 ± 400 850 ± 70 3000 ± 200 1400 ± 300 Vole-weasel 0.08 ± 0.01 0.025 ± 0.002 (10 ± 2) · 104 100 ± 50 200 ± 200 510 ± 60 1600 ± 200 730 ± 90 Wolf-moose 33 ± 1 350 ± 10 - - 40 ± 20 42 ± 2 - - Lemming osc. - 0.064 ± 0.003 1000 ± 200 14 ± 5 - - 1460 ± 0 860 ± 80

TABLE I. Table of the data used and the values calculated, including uncertainties. Numbers are rounded to the highest uncertain digit [16–22]. more rounded shape of the peaks, similar to the one seen lupus) and moose (alces alces americanus) of Isle Royale in actual observations. It turns out that the time average in Lake Superior, Michigan. On this island, wolves and abundances are fairly close to the predicted equilibrium moose coexist in isolation, with very little interference abundances in all of our numerical solutions. We shall from other animals. Due to the small size of the island, therefore use mean abundances and equilibrium abun- animal populations are so small that random events (such dances interchangeably when validating our results. as the introduction of parvovirus to the wolf population We also plug the masses into eqs. (14) and (17). The in 1980) will have a large influence on the population, theoretical estimates obtained this way and their uncer- which seems to fluctuate almost erratically [27]. There- tainties can be seen in table I. For this particular system, fore, we cannot determine an observational population we estimate a period of T 1400 days. Compared to the cycle length for this system. However, the average popu- observed period of around≈ 3000 days, the error is about lations should reflect an equilibrium ratio that should be 50 %. As far as order-of-magnitude estimates go, this is predictable from wolf and moose mass. As can be seen still reasonable. Neglecting the duration of the collapse in table I, we obtain an accurate estimate of this ratio. phase is probably part of the reason for this error. For Based on these cases we may conclude that our ide- comparison, the cycle period obtained from linearisation alised model works as an order-of-magnitude estimate of 2π 550 gives us = 1/8 = 770 days, which is far too the behaviour of ecosystems. There is a discrepancy be- √αδ (mxmy ) short. This again underlines the usefulness of approx- tween the derived period of population oscillations and imating the cycle as a series of instantaneous collapse what is observed, and this discrepancy cannot be ex- phases followed by exponential growth phases. plained entirely by experimental uncertainty. However, Another case where the basic Lotka-Volterra equa- our results reproduce two patterns observed empirically, tion might be useful is the interaction between the vole which have not yet been theoretically explained. One (microtus agrestis) and least weasel (mustela nivalis) in is the apparent scaling of oscillation periods with mass northern Scandinavia, as mentioned above. Although to the quarter power. Another is the scaling of period there still is some doubt about the role of predation in the with ln xmax . We will therefore argue that the derived xmin cycle here as well, there is evidence that predation plays expression is of interest despite the discrepancy. at least a significant part. Vole density ranges from 102 4 2 to 10 km− over a cycle, while weasel density ranges 2 from 1 to 20 km− and is strongly correlated with vole density at northern latitudes [16]. The cycle is observed V. DISCUSSION to be about 4 years long in the areas we are interested in [11]. A numerical solution of the Lotka-Volterra equa- As predicted in the introduction, we have been able to tions for these parameter values can be seen in fig. 2 (b). parameterise the Lotka-Volterra equations using animal This numerical solution gives us an estimate of the pe- body mass in such a way that they provide fairly accu- riod T 830 days = 2.3 years, and of the prey-predator rate predictions of the equilibrium predator-prey popu- ratio of≈ 460 voles per weasel. A comparison between the- lation ratio. When we also know the amplitude of the oretical results calculated using the derived expressions fluctuations of prey population, we obtain analytical es- and observations can again be seen in table I. timates of the oscillation periods that reproduce the pat- Large population oscillations are observed in some ro- terns found in nature, albeit with a discrepancy. Notably, dent species even when there is no single obvious predator our approximate expression for the cycle period exhibits feeding on the rodent. One example of this is the north- the same allometric mass scaling as the one found em- ern collared lemming of Greenland (dicrostonyx groen- pirically. Furthermore, it shows a logarithmic scaling of landicus) [17]. We of course cannot use such an example period with the ratio of maximum to minimum popula- to test our hypothesis about prey-predator population tions, which is also found in data. The ratios of average ratios, but we may still use it to examine the accuracy of prey population to average predator population found in the derived period. The results of our examination can our simulations fit relatively well with real-world data. be seen in the table, and both the estimated period and For the population ratios, the uncertainty of population the error are similar to those of the vole. counts and animal weights explain the errors in two of As a final example, we will consider the wolves (canis three cases. Our prediction of the amplitude of predator 7

2000 3500 Observations Observations 1800 Theory Theory 3000 1/4 (21/ )*(m /m )3/4 400*ln(x /x )*m 1600 y x max min x 2 / ( * ) ; m = 0.08 kg = 10 % y 2500 1400 = 5 % 2 / ( * ) ; m = 11 kg y = 20 % 1200 2000 1000 x/y

T [days] 1500 800

600 1000

400 500 200

0 0 0 1 2 3 4 5 6 7 8 10-2 10-1 100 m /m m [kg] (a) y x (b) x

FIG. 3. (Colour online) (a) shows the observed and theoretically calculated prey-predator population ratios for the wolf-moose, vole-weasel, and hare-lynx systems. Here, the full line shows the predicted power law. The dotted line shows the theoretically calculated prey-predator ratio for half the ecological efficiency used in this paper (5 %), while the dashed line shows the prey-predator ratio calculated for twice the used ecological efficiency (20 %). (b) shows the observed and calculated periods of population oscillations for voles, lemmings, and hares. The black line shows the corresponding mass power law where we have set ln xmax = 4. This number is close to the values for the vole and lemming oscillation. The dotted and dashed lines  xmin  1/8 show the (mxmy) scaling law predicted from linearisation, where my is that of lynx and weasel respectively. We could not calculate an oscillation period for the moose of Isle Royale, and no single predator is known to cause the lemming oscillation, so they each only occur in one of the plots. Data points and error bars show the numbers without any rounding. oscillations, however, is unreasonable in comparison with and that prey may survive for longer in some locations observations, possibly because of the assumption that the than in others may serve to slow the dynamics of real, predator is entirely dependent on one prey species. geographically extended ecosystems. This, however, is a Of course, even though our model was only meant as subject that we will leave for future studies. a crude estimate, we need to address why we see the dis- Despite these discrepancies, our work demonstrates crepancy that we do between theory and observations. In that, by using the many available allometric mass scaling the case of the prey-predator population ratios, the un- laws, it is possible to obtain reasonable predictions from certain estimate of ecological efficiency is a likely source even very simple population dynamical models. This fact of error. The range of ecological efficiencies observed in should have wide applications in population dynamics. the real world is so large that it poses a challenge to this Another area where this could be applicable is in epidemi- kind of population dynamical modelling. If our estimated ology. The incubation and recovery times of a variety efficiency is a bit too low, it will explain the discrepancy, of diseases with multiple host species have already been as can be seen in fig. 3. shown to scale with host mass [28], and Dobson [29] has The period is off by a larger percentage, and it is less studied a multi-host disease model parameterised using clear what might cause the error. One drastic assumption mass. A possible further use of the model described here that we have made is that it takes no time for animals to could be to construct an epidemiological model taking grow to adult size. We have considered whether this de- predation into account. Models of epidemics in predator- lay might explain some of the error. To take the time re- prey systems have been proposed before [30], but they quired to reach full size into account, we have attempted often contain so many unknown parameters that an ex- a numerical solution of the equations while including a amination of parameter space becomes difficult. Here, time delay in predator and prey reproduction. Unfortu- a parameterisation using mass could significantly reduce nately, this does not significantly change the oscillation the number of free parameters. period. Another possible source of error is the assump- Already in 1992, Yodzis & Innes pointed out that the tion that collapse is instantaneous. In reality, it does application of mass scaling relations to population dy- take some time, though not as long as the exponential namics can potentially make it a lot easier to make re- growth phase. If the duration of the collapse phase also alistic estimates of the parameters involved. Still, to scales with animal mass, it would help explain why we our knowledge it is not until now that the predictions of consistently underestimate the period by about 50 %. a mass-parameterised population dynamical model have Finally, our model is a mean-field theory, whereas in been tested against real-world data. The scaling of re- reality, geographical separation does play a role. Maybe production rate with animal mass has also provided us the fact that real predators have to seek out the prey, with a possible explanation for the relationship between 8 population cycle length and mass, at least in systems potentially prove highly useful in population dynamics. where the amplitude is large. This is for example very much the case for several rodent and lagomorph species. In conclusion, we find that the allometric mass scaling ACKNOWLEDGMENTS laws that apply to a variety of biological quantities could The authors wish to thank Dr. David Vasseur for his detailed and useful comments on an earlier draft of this article.

[1] Lev R. Ginzburg and Mark Colyvan, Ecological orbits: tine cycles in the northern fennoscandian taiga?” Oikos how planets move and populations grow (Oxford Univer- 50, 353–365 (1987). sity Press, 2004). [19] Michael Carlsen, Jens Lodal, Herwig Leirs, and [2] Robert Henry Peters, The ecological implications of body Thomas Secher Jensen, “The effect of predation risk on size (Cambridge University Press, Cambridge, 1983). body weight in the field vole, microtus agrestis,” Oikos [3] Geoffrey B. West, James H. Brown, and Brian J. En- 87, 277–285 (1999). quist, “A general model for the origin of allometric scal- [20] Ron Moen, James M. Rasmussen, Christopher L. Bur- ing laws in biology,” Science 276, 122–126 (1997). dett, and Katharine M. Pelican, “Hematology, serum [4] R.O. Peterson, R.E. Page, and K.M. Dodge, “Wolves, chemistry, and body mass of free-ranging and captive moose, and the allometery of population cycles,” Science canada lynx in minnesota,” Journal of Wildlife Diseases 224, 1350–1352 (1984). 46, 13–22 (2010). [5] P. Yodzis and S. Innes, “Body size and consumer-resource [21] Bruce J. Gillingham, “Meal size and feeding rate in the dynamics,” The American Naturalist 139, 1151–1175 least weasel (mustela nivalis),” Journal of Mammalogy (1992). 65, 517–519 (1984). [6] Ulrich Brose, “Body-mass constraints on foraging be- [22] U. S. Seal and L. D. Mech, “Blood indicators of seasonal haviour determine population and food-web dynamics,” metabolic patterns in captive adult gray wolves,” The Functional Ecology 24, 28–34 (2010). Journal of Wildlife Management 47, 704–715 (1983). [7] Alfred J. Lotka, “Analytical note on certain rhythmic [23] Charles J. Krebs, Rudy Boonstra, Stan Boutin, and relations in organic systems,” PNAS 6, 410–415 (1920). A.R.E. Sinclair, “What drives the 10-year cycle of snow- [8] Raymond Lindeman, “The trophic-dynamic aspect of shoe hares?” BioScience 51, 25–35 (2001). ecology,” Ecology 23 (1942). [24] Charles J. Krebs, “Of lemmings and snowshoe hares: the [9] R. Trebilco, J. K. Baum, A. K. Salomon, and N. K. ecology of northern canada,” Proceedings of the Royal Dulvy, “Ecosystem ecology: size-based constraints on the Society B 278, 481–489 (2011). pyramids of life,” Trends In Ecology & Evolution 28, [25] Garth Mowat, Mark O ’ Donoghue, and Kim Poole, 423–431 (2013). “Ecology of lynx in northern canada and alaska,” in Ecol- [10] P. A. Colinvaux and B. D. Barnett, “Lindeman and the ogy and conservation of lynx in the United States, edited ecological efficiency of wolves,” The American Naturalist by L.F. Ruggiero, K.B. Aubry, S.W. Buskirk, G. M. 114, 707–718 (1979). Koehler, C. J. Krebs, K.S. McKelvey, and J.R. Squires [11] Ilkka Hanski, Heikki Henttonen, Erkki Korpim¨aki, Lauri (Boulder: University of Colorado Press, 2000) Chap. 9, Oksanen, and Peter Turchin, “Smallrodent dynamics pp. 265–306. and predation,” Ecology 82, 1505–1520 (2001). [26] Ronald L. Smith, Dennis J. Hubartt, and Russell L. [12] C. J. Krebs, S. Boutin, R. Boonstra, A. R. Sinclair, J. N. Shoemaker, “Seasonal changes in weight, cecal length, Smith, M. R. Dale, K. Martin, and R. Turkington, “Im- and pancreatic function of snowshoe hares,” The Journal pact of food and predation on the snowshoe hare cycle,” of Wildlife Management 44, 719–724 (1980). Science (New York, N.Y.) 269 (1995). [27] John A. Vucetich and Rolf O. Peterson, “The influence of [13] S. Rinaldi and S. Muratori, “Slow-fast limit cycles in prey consumption and demographic stochasticity on pop- predator-prey models,” Ecological Modelling 61, 287–308 ulation growth rate of isle royale wolves (canis lupus),” (1992). Oikos 107, 309–320 (2004). [14] Melvin E Sunquist, Wild cats of the world (University of [28] Jessica M. Cable, Brian J. Enquist, and Melanie E. Chicago Press, Chicago, Ill, 2002). Moses, “The allometry of host-pathogen interactions (al- [15] P. Turchin, L. Oksanen, P. Ekerholm, T. Oksanen, and lometry and disease),” PLoS ONE 2 (2007). H. Henttonen, “Are lemmings prey or predators?” Nature [29] A. Dobson, “Population dynamics of pathogens with mul- 405 (2000). tiple host species,” The American Naturalist 164, 64–78 [16] I. Hanski, L. Hansson, and H. Henttonen, “Specialist (2004). predators, generalist predators, and the microtine rodent [30] Ying-Hen Hsieh and Chin-Kuei Hsiao, “Predator-prey cycle,” The Journal of Animal Ecology 60 (1991). model with disease infection in both populations,” Math- [17] Olivier Gilg, Ilkka Hanski, and Benoˆıt Sittler, “Cyclic ematical Medicine and Biology: A Journal of the IMA dynamics in a simple vertebrate predator-prey commu- 25, 247–266 (2008). nity,” Science (New York, N.Y.) 302 (2003). [18] Heikki Henttonen, Tarja Oksanen, Aarre Jortikka, and Voitto Haukisalmi, “How much do weasels shape micro-