Biology Reference
In-Depth Information
efficiency, Kacelnik generated the curve for them by dropping mealworms at successively
longer and longer intervals. The trained bird would simply wait on the wooden tray for
the next worm to arrive, until eventually it flew home with a beak-load for its chicks.
The beauty of this experimental method is that Kacelnik knew the shape of the loading
curve precisely, and was hence able to present exactly the same loading curve at
randomly varying distances (ranging from 8 to 600 m) from the nest on different days.
The results were striking (Fig. 3.2c): not only did the load size increase with distance
from the feeder to the nest, but there was a close quantitative correspondence between
the observed load sizes and those predicted by the model of maximizing delivery rate.
Let us briefly summarize what the results of the starling study show. We started off by
considering load size from the point of view of costs and benefits. We formulated a
specific hypothesis about how costs and benefits might influence load size in the form of
a model (Fig. 3.2a), and then used the model to generate a quantitative prediction (Fig.
3.2b). In making the model we did three important things. Firstly, we expressed a
general conviction that starlings are designed by natural selection to be good at their job
of parenting. This is not something that we aimed to test, but it is our general background
assumption to justify thinking in terms of maximizing pay-off in relation to costs and
benefits. Secondly, we made a guess about the
currency
of costs and benefits; we
suggested that for a parent starling the crucial feature of doing a good job is maximizing
net rate of delivery of food to the nestlings. Thirdly, we specified certain
constraints
on
the starling's behaviour. Some of these constraints are to do with features of the
environment (the time required to travel, the shape of the loading curve). Another
important assumption about constraints is that the starling is assumed to 'know', or at
least to behave as if it knows, the travel time and the shape of the loading curve. When
we worked out the optimum load size we assumed that these were known. The
experimental results supported the predictions of the model, and in so doing they
supported the hypotheses about the currency and the constraints that were used to
construct the model. Kacelnik compared the predictions of models based on several
different currencies and he found, for example, that one based on energetic efficiency
(energy gained/energy spent) as opposed to rate gave a rather poor fit to the data.
Box 3.1 shows that the same economic model we have used for the starlings can
be applied to other situations where individuals experience diminishing returns from
a patch.
A field test with
starlings
Optimality models
include
assumptions
about currencies
and constraints
Bees
A similar problem is faced by a worker honeybee as it flies from flower to flower filling its
honey crop with nectar to take back to the hive. Bees also often return to the hive with
less than the maximum load they could carry and their behaviour can be explained by
a model similar to that used for the starling. There is, however, an important difference:
the bee experiences a curve of diminishing returns neither because the nectar in its crop
makes it less able to suck more flowers nor because of resource depression (Box 3.1) but
because the weight of nectar in the crop adds an appreciable energetic cost to flight. The
more the bee loads up its crop the more of its load it will burn up as fuel before it gets
home. As a consequence, while the gross quantity of nectar harvested increases at a
constant rate, the
net
yield of energy for the hive increases at a diminishing rate as the
crop fills (producing, in effect, a loading curve like that of the starling).
