Hi Irvine
It could equally be my understanding of OFT that's at fault. Like most of these things, depending on who you read the same terms mean different things.
Not because it has consciously chosen to do so but because evolution has selected for the animals whose inherited set of reactions best suit the environment it occupies. As Paul says it’s a nebulous instinct.
The idea of a "fixed action plan" is an automatic, unthought, innate behavior, an instinctive behaviour released by environmental stimuli which must be carried through to its conclusion. Nothing nebulous about it. Fitness for an eco-niche means the total of decisions an organism takes, the actions it makes, must keep it alive and well and be sufficient to allow it to mature and breed. Nothing that nebulous there either.
Thing is there's a difference between those, behaviour is to do with specifics, it answers a proximate question. The sort of thing might be 'Does a fish choose a mate and if they do how do they choose?" or 'What causes an insect to hatch or delay hatching?"
There's a difference in type between those questions and things like "Why does natural selection favour insects hatching that way? In what way fish are who don't select an individual mate more fit?" - so-called ultimate questions. (No the answer isn't 42
)
If you ask "How does this insect forage? (find and handle food)" and answer "They forage optimally." Then I think you're asking a question of the proximate type and getting an answer to the other type of question. When I ask how a trout finds and handles food when it hunts, I want to know how it locates it, what factors cause or allow it to perceive that thing as prey and behave in an appropriate manner. If the answer to that is "They do it optimally." how useful is that? At worst it's redundant since all behaviour by a successful animal is optimal in some sense. At best, in that situation, acting optimally seems to imply trout has some superior knowledge of prey density in its immediate area, or even that it knows which bugs will be hatching in droves tomorrow.
"Optimal models often make precise predictions, e.g. switch to less preferred food when preferred food fall below set rate. In practice animals will sample less preferred option even at higher rates. Again modification to original model make it possible to predict partial preference." quoted from an online piece.
Optimal models predict trout hunting in the here and now will fully and optimally exploit an abundant hatch - reduce the time needed for encountering, identifying and handling prey to a minimum - in short they gorge on the most abundant bugs. The reality is they don't become absolutely fixated on the most abundant hatch. Trout also 'sample' other bugs.
In my fishing experience, when trout are feeding hard on an abundance of one prey its often a very good tactic to offer something that looks nothing like the caenis of mayfly or buzzers or.... The way I understand that, optimal applies to why they do what they do. Trout food varies a lot through the season, its in the best interests of their fitness that they sample other prey. That means how they behave may appear sub-optimal - they waste time and energy nailing a daddy when there are thousands of buzzers in the water. But optimal in the ultimate sense that they can switch their attention to other prey with a minimum of adjustment.
Incidentally if you think about how trout 'in the zone', gorging on an abundant hatch, behave and how they recognize prey it has interesting implications for us. In my experience and in the experience of many other anglers I know, fish feeding hard are more difficult to spook, far less 'wary' far easier to approach. They appear to identify their prey using a minimum of cues, patterns can be more general, things like size and simple encounter may be the deciding factors. Far less shilly-shally, if they meet it they eat it or ignore it. Takes tend to be calm, they just suck it in and look for the next.
Magnus