There are four principal empirical ways to assess plant responses to temperature:
(1) looking into the past, using historical trends of temperature and growth, in
essence restricted to dendrology, (2) studying current growth processes across thermal
gradients, or (3) studying current growth in response to the natural temporal
variation in temperature and (4) manipulating temperatures around plants and testing
their responses under controlled conditions (both indoors and in the field). Each
of these approaches has some advantages and disadvantages. While type 4 tests
are best controlled in terms of environmental influences, they are limited in time
and space and are commonly confined to very artificial growth conditions and obviously
restricted to very young ages in the case of trees. The other three options
are commonly less ‘precise’ in the sense of isolating temperature effects from other
effects and good replication, but they are closer to real world conditions. It is the
challenge of empirical sciences to make maximum use of all these options, but there
is a great need to complement the predominance of type 4 studies with more type
1–3 studies (K¨orner, 2001). The area second best explored is tree rings that, for
instance, allowed the demonstration of clear warming effects in treeline trees in
recent decades (Rolland et al., 1998; Paulsen et al., 2000) in some regions but not
in others (Kirchhefer, 2005). I would like to argue for greater attention to type 2 and
3 studies.