Biology Reference
In-Depth Information
The physical or biological environment may fluctuate too rapidly for the
animals to 'catch up' in their adaptations (Chapter 4). We provided an example
of this in Chapter 1, when discussing how some populations of birds have not
advanced their breeding times sufficiently to keep up with the rapid
advancement of spring in recent decades.
(i)
… evolutionary
lag …
Adaptation could be constrained by the underlying genetics of the behaviour.
One reason is that there may be insufficient genetic variation for new strategies
to evolve. If the environment changes or if for some other reasons the optimal
phenotype changes, animals can adapt to the new conditions only if there is
genetic variation in the population. However, this is unlikely to be of general
importance, because whenever genetic variation is looked for, it is usually
found (Lynch & Walsh, 1998).
(ii)
… genetic
constraints …
Another reason is that traits could be genetically linked, in which case
selection on one trait could influence the other. An example of this is provided
from work on Soay sheep living on the island of St Kilda, where coat colour is
either dark brown or light tawny (Fig. 15.3). Variation in coat colour is
controlled by a single locus called
TYRP1
, where the dark allele (G) is dominant:
GG and GT give a dark coat, whereas TT gives a light coat. Over a 20-year
period the proportion of sheep with dark coats and the G allele has been
decreasing (Fig. 15.3). This cannot be due to selection for cryptic colouration
(Chapter 4) or sexual selection (Chapter 9) because there are no predators on
the island and mate choice does not depend upon coat colour.
Instead, Gratten
et al
. (2008) found that the decline in frequency of dark
coats is due to selection on some other gene (or genes) for fitness that are
physically close to
TRYP1
and which cause homozygous GG sheep to have a
reduced fitness, relative to both GT and TT. The linked genes and why they
affect fitness have yet to be identified. This study documents an unusual case,
because the lack of predators and coat dependent mate choice means that coat
colour
per se
is not under selection, and so it is free to be pulled along by linked
genes. Presumably, if this was not the case, and there was a selective advantage
to dark coats, then there would be selection to break the association between
coat colour and the unknown fitness effect. Consequently, whilst the underlying
genetics can constrain adaptation, this does not necessarily lead to a dead end,
because the underlying genetics themselves are also subject to selection.
The way in which ESS models ignore the underlying genetics and assume all
phenotypes are possible has been termed the 'phenotypic gambit' by Alan
Grafen (1984). On one hand, the phenotypic gambit can be seen as a pragmatic
approach, which makes it easier to develop models. However, it can also be
argued that it is often the most useful approach. Working out an exact solution
to a specific genetic model may not be very helpful, because there will be a
'cloud' of possible models, which differ in the genetic details (about which we
are unlikely to ever know). The phenotypic gambit provides a robust
approximation to a wide range of models, which can therefore tell us more
than the exact solution of a more specific genetic model. The key point here is
that the phenotypic gambit delivers approximate predictions that can be
