5. Adaptive Radiation
81
al. 1995). By comparison, only 5% of the historically known species of Clermontia
are extinct, and 13% considered endangered. The principal factors associated with
the likelihood of historically known species becoming extinct appear to be (i) initial
rarity (see Terborgh and Winter 1980); (ii) occurrence in areas heavily disturbed by
humans; and (iii) possession of highly specialized flowers > 45 mm in length
(Givnish et al. 1995). Ten of the 13 extinct species of Cyanea were known from
only one site, compared with only 4 of 52 extant species (X2 = 27.1, P < 0.0001 for
1 d.f.). To the extent that data are available, the average elevational breadth of the
extinct species was significantly less than that for extant species (102 m vs. 424 m,
P < 0.005, 2-tailed t-test with 51 d.f.). Finally, the extinct species C. arborea, C.
camata, C. pohaku, and C. quercifolia all occurred on leeward West Maui, in habitats that were heavily logged and/or cleared for pastures or Eucalyptus plantations;
C. giffardii occurred on a single kipuka near Glenwood on Hawai'i that was largely
cleared for cattle grazing (Givnish et al. 1995).
3 Limited Dispersal and Convergent Adaptive
Radiations as a General Pattern
The general principle that limited dispersal can interact with selection for ecological divergence to produce multiple, parallel adaptive radiations finds support in
several other groups. Examples include the paraliel radiations of marsupial and
placental mammals into a wide range of ecological roles on Australia vs. other
continents (Springer et al. 1997); of Analis lizards into a set of ecomorphs that
partition the same set of structural habitats on each island of the Greater Antilles
(Jackman et al. 1997; Losos et al. 1998); of arboreal, semi-arboreal, and terrestrial
forms in different lineages of Mandarina land snails, on various islands of the
Bonin chain since the Pleistocene (Chiba 1999); and of three floral syndromes in
three habitat classes by different clades of the bulbous, heavy-seeded plant genus
Calochortus (Patterson 1998)". Each of these findings belies the claim by Gould
(1989) that a "replay of the tape [of life] would lead evolution down a pathway
radically different from the road actually taken". They do not support Gould's assertion that historical contingency - the particular lineage undergoing radiation,
its genetic and phylogenetic constraints, and the suite of competitors, predators,
and mutualists it faces - plays a predominant role in determining the direction of
evolution. It must be recognized, however, that evolution within each of these groups
has involved diversification based on very similar patterns of development and
genetic variation (see Givnish 1997). Historical contingency may be important in
determining which lineages undergo adaptive radiation in particular contexts, but
ecological factors may be crucial in determining the pattern of adaptive radiation
and species diversification that results.
81
al. 1995). By comparison, only 5% of the historically known species of Clermontia
are extinct, and 13% considered endangered. The principal factors associated with
the likelihood of historically known species becoming extinct appear to be (i) initial
rarity (see Terborgh and Winter 1980); (ii) occurrence in areas heavily disturbed by
humans; and (iii) possession of highly specialized flowers > 45 mm in length
(Givnish et al. 1995). Ten of the 13 extinct species of Cyanea were known from
only one site, compared with only 4 of 52 extant species (X2 = 27.1, P < 0.0001 for
1 d.f.). To the extent that data are available, the average elevational breadth of the
extinct species was significantly less than that for extant species (102 m vs. 424 m,
P < 0.005, 2-tailed t-test with 51 d.f.). Finally, the extinct species C. arborea, C.
camata, C. pohaku, and C. quercifolia all occurred on leeward West Maui, in habitats that were heavily logged and/or cleared for pastures or Eucalyptus plantations;
C. giffardii occurred on a single kipuka near Glenwood on Hawai'i that was largely
cleared for cattle grazing (Givnish et al. 1995).
3 Limited Dispersal and Convergent Adaptive
Radiations as a General Pattern
The general principle that limited dispersal can interact with selection for ecological divergence to produce multiple, parallel adaptive radiations finds support in
several other groups. Examples include the paraliel radiations of marsupial and
placental mammals into a wide range of ecological roles on Australia vs. other
continents (Springer et al. 1997); of Analis lizards into a set of ecomorphs that
partition the same set of structural habitats on each island of the Greater Antilles
(Jackman et al. 1997; Losos et al. 1998); of arboreal, semi-arboreal, and terrestrial
forms in different lineages of Mandarina land snails, on various islands of the
Bonin chain since the Pleistocene (Chiba 1999); and of three floral syndromes in
three habitat classes by different clades of the bulbous, heavy-seeded plant genus
Calochortus (Patterson 1998)". Each of these findings belies the claim by Gould
(1989) that a "replay of the tape [of life] would lead evolution down a pathway
radically different from the road actually taken". They do not support Gould's assertion that historical contingency - the particular lineage undergoing radiation,
its genetic and phylogenetic constraints, and the suite of competitors, predators,
and mutualists it faces - plays a predominant role in determining the direction of
evolution. It must be recognized, however, that evolution within each of these groups
has involved diversification based on very similar patterns of development and
genetic variation (see Givnish 1997). Historical contingency may be important in
determining which lineages undergo adaptive radiation in particular contexts, but
ecological factors may be crucial in determining the pattern of adaptive radiation
and species diversification that results.