Technical Comments

Response to Comment on "Buffered Tree Population Changes in a Quaternary Refugium: Evolutionary Implications"

See allHide authors and affiliations

Science  07 Feb 2003:
Vol. 299, Issue 5608, pp. 825
DOI: 10.1126/science.1080630

Stewart (1) reiterates recently published views on the possible existence of “cryptic” refugia in northern Europe (2) and questions the role of southern European refugia as centers for speciation. In view of the implications of these issues for conservation strategies, critical evaluation of the genetic and paleobotanical evidence is particularly important.

Establishing the northernmost extent of glacial refugia in Europe is not straightforward, because a species' distribution pattern is likely to comprise larger southern populations and progressively smaller, northern peripheral ones (3). Moreover, postglacial expansion may include advances and extirpations due to climatic oscillations, and thus northern remnants of an earlier advance may form the “secondary” refugia for a subsequent wave of expansion (4). Separating the effect of these events and resolving the location of the northern edge during glacial maxima requires spatially detailed genetic and fossil evidence, which are still largely unavailable. It is also important to consider the individual histories of species: Making inferences about refugia of temperate trees based on evidence from freshwater fish, land snails, sedges, or rockferns may be of limited value.

Most paleoecologists accept that small populations of conifers and birch survived north of the Alps during the Last Glacial Maximum (LGM). Stewart (1) challenges the dominant paradigm by invoking the occurrence of refugia of temperate trees with seriously disjunct distributions north of the Iberian, Italian, and Balkan peninsulas. The impetus for this claim originates from the existence of fossils of temperate mammals—which today are associated with deciduous woodlands—in glacial deposits in northern Europe. We question the strength of this association because these mammals are found in a variety of habitats including open, montane, and coniferous environments (5, 6). In addition, no clear genetic evidence exists for disjunct refugia far north of the main refugial range of temperate trees in Europe. Finally, the paleobotanical evidence cited by Stewart (2) does not stand up to close scrutiny. The occurrences of temperate trees in (2) were either not LGM in age (19 to 23 ka), or were dated indirectly and their chronological framework is unclear.

We suggest that the existence of northern refugia can be tested using continuous pollen records spanning the last glacial north of the Alps. If small populations of temperate deciduous trees survived in these regions, they should have been able to expand during the 2000- to 3000-year interstadial phases that punctuated the period between 45 and 11.5 ka. No such increase is evident from French pollen records, which only show expansion of pine and birch and some spruce and larch (7, 8). Rather, the available evidence indicates that postglacial colonization by temperate trees did not originate north of about 45°N and that their glacial survival and migration histories may have differed from the examples cited in (1). Where that leaves temperate mammals with northern European glacial records is unclear, but behavioral accommodation and facultative adaptation (9) may be easier to invoke than trying to force the paleobotanical record to fit the mammalian story.

The second issue raised in (1) is that southern European refugia of temperate trees could not be both sources for northward postglacial colonization and crucibles of speciation, because they could not remain sufficiently isolated for genetic divergence to accumulate. We suggest that the two roles are not incompatible. The history of a refugium can be viewed as the continuous presence of tree populations with intermittent pulses of outward range expansion. Encounters with other expanding genomes from different geographical areas would take place farther afield and lead to the formation of hybrid zones (10). The keys to long-term isolation are that (i) a part of the refugial population remains behind providing continuity; and (ii) during intervals of climatic deterioration, the colonizing genomes do not return to their point of origin, but rather degrade in situ (11). Whether the refugial populations we identified in the Pindus Mountains (12) contributed to the recolonization of Greece but were blocked in the north by a leading edge expansion of Balkan populations (10), or also contributed to northward expansion beyond Greece, is irrelevant to the issue of long-term isolation. In either case, there was no reverse movement during the population contraction phase at the onset of a stadial or a glacial period, thus preserving the distinctive character of the tree populations that persisted in refugia. Any genetic differentiation that may have accumulated in northern European temperate tree populations during interglacials was ultimately lost in the ensuing glacial phase (11).


Stay Connected to Science

Navigate This Article