Contribution
Extends the hermit-crab trade-off paradigm (LIT-1450, LIT-1470) to an insect. It adds a control the crab studies lacked: the choice runs on conditioned cues rather than on contact with the stimuli, so the trade-off is made in the brain.
Key insight
How far a bee avoids a noxious feeder depends on what the alternative offers. “Bees avoided noxious feeders when the unheated feeders contained high sucrose concentrations, but progressively increased feeding from noxious feeders when the sucrose concentration at unheated feeders decreased” (p. 1).
Assumptions
- That 55 °C is noxious to a bee. It is supported by the drop in feeding when both offers are equal (0.5 to 0.3).
- That a context-dependent, centrally mediated reduction of nocifensive behaviour is evidence bearing on pain. The authors present it as a criterion from the animal-pain literature (refs 12–14), and say “it is not a formal proof” (p. 3).
Key results
- Without heat, bees preferred 40% feeders to 10–30% ones (p. 1).
- With heat and equal 40%/40% offers, they avoided the heated feeders. The mean proportion of feeding on the heatable feeder went from 0.5 to 0.3 (p. 1).
- In both conditions, feeding on the 40% feeders fell as the alternatives got richer, and more steeply when the 40% feeders were heated (p. 2). The corrected statistic is z = −2.132, P = 0.033.
- Landings without feeding fell from 2.45 in the first bout to 1.09 in the last, so choice ran on learned colour and position cues (p. 2).
From the deposited data, as the mean per-bee proportion of feeds on the 40% feeders (unheated → heated):
| alternative | n | unheated | heated |
|---|---|---|---|
| 10% | 10 | 0.997 | 1.000 |
| 20% | 10 | 0.964 | 0.817 |
| 30% | 2 | 0.782 | 0.653 |
| 40% | 10 | 0.467 | 0.279 |
The 30% group has two bees; the paper’s n = 32 is 10 + 10 + 2 + 10.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Bees trade off noxious-heat avoidance against sucrose preference | strong | the heat × concentration effect (corrected z = −2.132, P = 0.033), n = 32 |
| C2 | The trade-off is centrally mediated | moderate | fewer exploratory landings over bouts, so choice by learned cues |
| C3 | The ability is consistent with a capacity for pain in insects | weak | inference from a criterion, which the authors flag is not proof |
Concepts
- motivational trade-off — flexible weighing of two competing motivations, here heat avoidance against sucrose preference.
- nocifensive behaviour — the response to a noxious stimulus; “modulation” is its context-dependent reduction.
Connections
Builds on the hermit-crab studies (LIT-1450, LIT-1470), which it describes as showing that “hermit crabs require higher voltages of electric shock to evacuate their preferred shell species than for a less preferred one (5, 6)” (p. 1). Its precautionary conclusion cites Birch on animal sentience and the precautionary principle (ref. 15), which is not in the record.
Bearing on the record
On THEORY-101’s test. The study is an instance of what THEORY-101’s promotion asks for, “an animal accepting a noxious cost for a resource”, and in an insect. It meets “in proportion to both” only halfway. The resource is varied in four steps, but the cost is fixed at 55 °C, so nothing here shows the cost a given resource will buy.
Continuous or thresholded? At the one cost tested, graded. Heated feeding rises steadily as the alternative gets poorer, from 0.28 against a 40% alternative to 0.82 against 20% and 1.00 against 10%. Nowhere does the heat override the resource: against a 10% alternative every bee fed only from the heated feeders. That is compensatory potential in CLAIM-435’s sense, at least for this cost. But it does not test the claim’s question. A level threshold, ALARM’s override (THEORY-090), would appear as a cost above which no sucrose difference brings the bee back. With one cost and that cost evidently below any such threshold, the study cannot see one. Nor can it see the Tversky-style cycles that a discrimination-threshold override would produce (LIT-1472). That needs several costs, choices made within one subject, and pairs chained so that cost steps are small while resource steps are large.
Limits on reading the curve. The four resource levels are run in separate groups, one group having two bees. The lines in Fig. 1B are a generalized linear model’s predictions, a smooth form that assumes the gradedness rather than testing for a step. The proportion is pooled over bouts. So “graded” here means the group means are monotone and show no step at four levels. It does not mean a continuous indifference curve has been fitted. The deposited data could be refitted with a step model against a smooth one, and a single cost level would still limit the result.
On potential. For CLAIM-440 the trade-off is a negotiation between a heat-avoidance ranking and a sucrose ranking, in which neither prevails outright. The heat’s influence grows as the sucrose difference shrinks. Where the two offers are equal it decides the bee’s feeding about two to one (0.28 against 0.72), and where the difference is large it decides nothing. On the claim’s reading that intensity is the strength of potential, this is a regime of weak, compensatory potential of the cost ranking. The record of which ranking each choice followed, which QUESTION-268 asks for, is here only at the level of group proportions.
Limitations
Brief report; one cost level; between-group resource levels, one group very small; the proportion pooled over bouts; the corrected figure not seen here; no test of memory or of later avoidance, the other marks Elwood’s criterion lists.
Open questions
Whether a range of temperatures would show the price of a given sucrose advantage rising smoothly with the cost, or a temperature above which no sucrose brings the bees back. And whether, past 55 °C, the cost begins to decide regardless of a small sucrose difference, the mark of an override set on differences.