Climate-Driven Range Expansion and Morphological Evolution in a Marine Gastropod

See allHide authors and affiliations

Science  01 Jun 2001:
Vol. 292, Issue 5522, pp. 1707-1710
DOI: 10.1126/science.1060102


Little is known about the phenotypic consequences of global climate change, despite the excellent Pleistocene fossil record of many taxa. We used morphological measurements from extant and Pleistocene populations of a marine gastropod (Acanthinucella spirata) in conjunction with mitochondrial DNA sequence variation from living populations to determine how populations responded phenotypically to Pleistocene climatic changes. Northern populations show little sequence variation as compared to southern populations, a pattern consistent with a recent northward range expansion. These recently recolonized northern populations also contain shell morphologies that are absent in extant southern populations and throughout the Pleistocene fossil record. Thus, contrary to traditional expectations that morphological evolution should occur largely within Pleistocene refugia, our data show that geographical range shifts in response to climatic change can lead to significant morphological evolution.

Pleistocene [1.8 million to 10,000 years before the present (yr B.P.)] climatic fluctuations caused dramatic shifts in the geographic distributions of many species, both terrestrial and marine (1, 2). The effects of these range fluctuations on genetic population structure and speciation have received much attention (3–6), but little is known about their phenotypic consequences. Here we investigate the effects of Pleistocene climatic changes on evolution in a Californian marine gastropod, Acanthinucella spirata(Blainville, 1832).

Fossiliferous late Pleistocene terraces are common along the coast of California and preserve over 77% of the shallow-water molluscan species still living in the region (7). This fossil record shows that the geographic distributions of many shallow-water Californian mollusks changed in response to late Pleistocene climatic shifts (2, 8, 9). The excellent preservation of mollusks in these terraces provides a unique opportunity for directly measuring morphological changes associated with the Pleistocene range shifts.

A. spirata is a common intertidal carnivorous gastropod that presently ranges from Tomales Bay, California (38.2°N), to Punta Baja, Baja California (29.9°N). Juveniles emerge directly from egg capsules attached to rocks (10), without an intervening planktonic larval stage; hence the dispersal ability of A. spirata is low and the potential for population differentiation is high. Previous work revealed a difference in shell morphology between Pleistocene and Recent populations, suggesting a climatically driven late Pleistocene recolonization of the northern part of the species' range from a southern refugium (11).

To test this phylogeographic prediction, we collected individuals from 14 populations of A. spirata between San Diego and Tomales Bay, a distance of over 1050 km (12). A 660–base pair fragment of the mitochondrial gene encoding cytochrome c oxidase subunit I (COI) was amplified and sequenced from 117 individuals (13). The 33 unique haplotypes (GenBank accession numbersAY017492-AY017524, AY027511-AY027516, andAY027687-AY027694) define three well-supported (parsimony and neighbor-joining bootstrap support >95%) clades: one to the north of Point Fermin (in southern Los Angeles County) and two to the south (Fig. 1). Only the Point Fermin population contains haplotypes from all three clades, supporting the hypothesis that the Los Angeles region may have served as a refugium during late Pleistocene cooling episodes (11). Genetic diversity (as measured by the mean number of pairwise differences between individuals) in the Point Fermin population is also more than twice as high as in any other population. The phylogeographic break between northern and southern clades (Fig. 1) coincides geographically with one previously reported in a teleost fish (14) but lies well south of the major biogeographic boundary at Point Conception (15–18).

Figure 1

Parsimony network for 33 unique mitochondrial COI haplotypes obtained from sampling 117 A. spirata from 14 populations from along the coast of California. Numbers in parentheses indicate the number of identical haplotypes from a locality. Open circles indicate inferred mutational steps between sampled haplotypes. Populations are as follows. North of Point Conception (shown in blue): Tomales Bay (TB), Half Moon Bay (HM), Monterey Bay (MO), San Simeon (SS), Cayucos (CY), and Jalama Bay (JB). South of Point Conception: El Capitan State Beach, Santa Barbara (SB); Leo Carillo State Beach (LC); Point Fermin (PF) and Cabrillo Beach (CP), both in San Pedro (shown in red); Carlsbad (CB); La Jolla (LJ); Mission Bay (MB); and Ocean Beach (OB).

The most striking pattern in the genetic data is a steep decline in genetic diversity from south to north (Fig. 1). Each of the six southernmost populations has a different most-frequent haplotype, but a single most-common haplotype is shared by all eight populations north of Point Fermin, which together span a geographic distance over four times greater than that covered by the six southernmost populations (Fig. 1). The decline in genetic diversity is even more dramatic across Point Conception. In populations of A. spirata north of this boundary, no haplotype other than the most common one occurs more than once (Fig. 1). Results from an analysis of molecular variance (AMOVA) (19) show that populations north of Point Conception are not subdivided, whereas among-population variation accounts for a large proportion (72%) of the total variance in populations south of Point Conception (P < 0.001, Table 1).

Table 1

Genetic variation and population subdivision north and south of Point Conception.

View this table:

The decline in nucleotide diversity north of Point Conception could result from either a recolonization of that part of the range from a southern refugium or from a selective sweep. For A. spirata, the range expansion hypothesis is supported by multiple lines of evidence as follows. (i) Phylogenetic rooting of A. spiratahaplotypes using two closely related species (20), A. paucilirata (GenBank accession number AY017490) and A. punctulata (GenBank accession number AY017491), as outgroups shows that the southern clades are ancestral to the northern clade (21). (ii) Pairwise haplotype differences in the populations north of Point Conception show a unimodal decline from a peak at zero differences, whereas the southern populations show a more ragged distribution (Fig. 2), a difference consistent with a recent and rapid population expansion north of Point Conception (22). (iii) Several other temperate marine and terrestrial species found in habitats restricted to the coast of western North America show the same pattern of reduced genetic variation toward the northern end of their ranges (23–25). Such parallel patterns in different genetic markers in taxa as disparate as birds and mollusks are more likely to stem from similar responses to past climatic changes in the region rather than from multiple independent selective sweeps. (iv) The sharp change in haplotype diversity occurs at Point Conception, a major biogeographic boundary marked by large changes in oceanographic conditions that have a strong influence on present-day species range limits (15,16, 18). (v) Specimens of A. spirata are present in middle Pleistocene deposits throughout California and are common in late Pleistocene deposits (oxygen isotopic stage 5e, about 125,000 yr B.P.) south of Point Conception (9), but they are either absent or extremely rare in late Pleistocene assemblages from the north (26). Taken together, these different lines of evidence support the range expansion hypothesis rather than a selective sweep. Furthermore, judging from the extreme homogeneity of northern populations (π = 0.45, no transversions), this expansion took place quite recently, around 12,000 to 31,000 yr B.P. (27).

Figure 2

Distributions of pairwise base pair differences in COI sequences between sampled A. spirata individuals north (left) and south (right) of Point Conception.

To investigate the phenotypic consequences of rapid northward range expansion, we measured the shape and size of shells from A. spirata individuals from living populations and from middle and late Pleistocene terrace assemblages (28). Individuals from all populations north of Point Conception combined differ significantly from southern individuals in both shell shape and size (Fig. 3 and Table 2). In terms of shape, the largest north-south difference is along the second principal component axis (PC2), which reflects differences in the relative height of the spire (Fig. 3 and Table 2). Shell shapes and sizes of Pleistocene individuals differ significantly from Recent northern populations combined but not from those to the south of Point Conception (Fig. 3 and Table 2). Most notably, the median PC2 scores for Recent northern populations fall outside the 95% confidence intervals for Pleistocene and for Recent southern populations combined (Fig. 3).

Figure 3

Spatial and temporal trends in shell shape and size in A. spirata. ( A) The line drawing on the left shows the 10 positional landmarks used in this study. On the right, the blue silhouette is for a northern specimen and the red is for a southern specimen. Drawings are not to scale. Note changes in the relative spire height as well as in the angularity of the whorls. (Bthrough D) Box plots of the scores on the first three principal component axes. The notches in each box show the 95% confidence interval around the median, and the top and bottom lines show the 90th and 10th percentiles, respectively. The blue boxes represent individuals from all populations north of Point Conception, the red boxes represent all individuals from populations to the south of this point, and the gray boxes represent all Pleistocene individuals. PC1 corresponds primarily to shell size; PC2 and PC3 correspond to various aspects of shape.

Table 2

MANOVA of the scores on the first three principal components axes summarizing morphological patterns.

View this table:

In combination, these genetic and morphological data not only indicate that there has been significant phenotypic evolution of A. spirata since the late Pleistocene, with changes concentrated in populations north of Point Conception, but also show that the genetically depauperate recolonized portion of the range is inhabited by morphotypes that are absent or rare in the fossil record of this species (Fig. 3). This significant difference in morphology between the Recent northern populations and the Pleistocene assemblages is unlikely to be due to sampling because (i) sample sizes for the Pleistocene and Recent populations are comparable (179 individuals versus 161 individuals, respectively), and (ii) we collected living individuals only from high intertidal habitats to control for the possibility of depth-related phenotypic variation, whereas the fossil assemblages are time-averaged and presumably drawn from a larger range of depths and microhabitats (2) that should inflate morphological variation. Thus, the novel northern morphotypes have either evolved since the expansion of this species' range or alternatively were rare in the past (and hence not preserved in the record) but are currently being favored in the northern part of the range of A. spirata. Thus, our data show that extinction-recolonization dynamics associated with climatic change have resulted in significant morphological changes in this species.

A species' response to climatic change can involve shifts in geographic range limits as well as adaptation to new conditions (2, 29, 30). Phenotypic responses to Pleistocene and Holocene climatic fluctuations have been documented in rodent populations (31, 32) as well as in insular land snails (33). The present results demonstrate that phenotypic changes can occur not only as an alternative to a geographical range shift but also in association with a range expansion.

Pronounced morphological change in a recolonized region contradicts the expectation that such changes should largely occur within Pleistocene refugia (34). In fact, A. spirata from the refugium (the San Pedro region) are unremarkable in terms of morphology. Instead, the picture revealed here accords with ecological studies in which rapid morphological evolution followed experimental (35) or anthropogenic (36) introductions into unoccupied habitat. Our data suggest that rapid morphological evolution may also be common in nature when changing climatic conditions alter the geographic ranges of species. Such a conclusion is further supported by data from freshwater fishes, in which post-Pleistocene colonization of new habitats has also led to evolutionary divergences (37, 38).

Ongoing studies of biotic responses to global change have either focused on ecological aspects such as changes in community compositions or species associations (1–4, 8) or on genetic changes (5, 6). The phenotypic consequences of global climate change have remained surprisingly neglected despite the excellent Pleistocene fossil record of many vertebrate and invertebrate taxa. Our data show that depending exclusively on genetic markers to judge evolutionary responses to environmental change may provide only an incomplete picture of biotic response. A better understanding of how species respond to climatic changes requires that we integrate ecological, genetic, morphological, and paleontological data.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: kroy{at}


View Abstract

Navigate This Article