Pike, nutrients and a green lake: reading a factorial enclosure experiment
In a eutrophic shallow lake (total phosphorus 95 µg/L), a team suspends sixteen 1 m³ enclosures, each sealed to the sediment. The design is factorial: pike present or absent, crossed with nutrient addition (N and P) or no addition, with four replicate enclosures per treatment. All enclosures start with the same water and plankton community. After eight weeks the means are:
| Treatment | Chlorophyll a (µg/L) | Zooplankton biomass (mg/L) | Secchi depth (m) |
|---|---|---|---|
| No pike, no nutrients | 48 | 0.9 | 1.1 |
| No pike, + nutrients | 96 | 0.5 | 0.6 |
| Pike, no nutrients | 22 | 2.6 | 2.6 |
| Pike, + nutrients | 78 | 0.7 | 1.0 |
Pike were given a fixed ration of minnows, so the pike themselves were never food-limited. The lake manager asks whether stocking pike could replace reducing phosphorus inputs from farmland.
1. Describe the interaction between the pike and nutrient treatments, and explain why reporting only the two 'main effects' would misrepresent these results.
Read the worked answer
Calculating the effects
- Effect of adding pike, at low nutrients: 48 → 22 µg/L, a 54% reduction in chlorophyll.
- Effect of adding pike, at high nutrients: 96 → 78 µg/L, a 19% reduction.
- Effect of adding nutrients, without pike: 48 → 96 µg/L, a doubling.
- Effect of adding nutrients, with pike: 22 → 78 µg/L, a 3.5-fold increase.
The interaction The strength of the pike effect depends on the nutrient level: pike suppress phytoplankton strongly when nutrients are low and only weakly when nutrients are high. The effects are not additive, so a main-effect statement such as 'pike reduce chlorophyll by an average of 37%' would be arithmetically true and biologically misleading - it would average over two situations that behave differently, and the manager would be given a number that applies to neither.
The anomaly worth flagging Zooplankton biomass with pike is 2.6 mg/L at low nutrients but only 0.7 mg/L at high nutrients. Pike presence is identical in both, and food for grazers is more plentiful in the second, so the drop cannot be a simple cascade. The data imply that nutrient addition changes the phytoplankton (for example towards inedible or colonial forms, or cyanobacteria) so that grazing is decoupled from algal biomass. Any explanation that ignores this does not fit the table.
2. Explain the mechanism the team is proposing and state two observations that would falsify it.
Read the worked answer
Proposed mechanism - a four-level trophic cascade
pike (top predator) → minnows (planktivores) → zooplankton (grazers) → phytoplankton.
- Adding pike reduces minnow numbers, either by direct predation or by forcing minnows into refuge habitat.
- Fewer minnows means lower predation on zooplankton, so zooplankton biomass rises (0.9 → 2.6 mg/L in the low-nutrient pair).
- Higher grazing pressure removes phytoplankton, so chlorophyll falls (48 → 22) and water clarity improves (Secchi 1.1 → 2.6 m).
- The cascade is top-down control; the nutrient treatment is bottom-up control, and the interaction shows that the bottom-up effect can overwhelm the top-down one.
Falsifying observations
- No decline in minnow biomass in the pike enclosures - then the pike effect on phytoplankton cannot be running through the planktivore link and must be direct or chemical.
- Pike added to enclosures that contain no minnows at all producing the same fall in chlorophyll - this would show pike act by excreting nutrients or by eating zooplankton directly, not via a cascade.
A further check is that a cascade requires the zooplankton increase and the chlorophyll decrease to occur in the same enclosures, not merely as treatment means.
3. Evaluate the manager's proposal to stock pike instead of reducing phosphorus input. Address the strength of the enclosure evidence and the ecological argument.
Read the worked answer
Strengths of the evidence It is a controlled, replicated, factorial experiment with the same starting water and plankton in every enclosure, so the contrast between treatments is internally valid and the interaction is real within the system studied.
Limitations that bear directly on the proposal
- Scale and duration: 1 m³ enclosures for eight weeks cannot reproduce a whole-lake food web. Pike are mobile, territorial predators with large home ranges; in a lake they will not distribute themselves evenly across the patches the manager wants treated. Wall effects, no fish immigration and no wind-driven sediment resuspension all differ from a real lake.
- Pike were fed a fixed ration, which removes the feedback that would occur in a lake, where pike would deplete minnows and then decline or switch prey.
- The residual chlorophyll is still unacceptable. Even with pike and added nutrients, chlorophyll is 78 µg/L and Secchi is 1.0 m - worse than the no-pike, no-nutrient control. The best treatment (22 µg/L, 2.6 m) only occurred when nutrients were also low.
- Mechanism mismatch: biomanipulation controls grazing pressure; it does not remove phosphorus. The sediment phosphorus store continues internal loading, and the data show that when nutrients are high the grazer control largely fails.
Conclusion Pike stocking may give a modest, temporary improvement in clarity and could be a useful supplementary measure where nutrients are already being addressed, but the experiment provides no support for using it as a replacement for reducing phosphorus inputs from farmland.