Cenozoic Insect–Plant Diversification in the Tropics
Total Page：16
File Type：pdf, Size：1020Kb
COMMENTARY Cenozoic insect–plant diversification in the tropics Donald R. Strong* and Michael Sanderson Section of Evolution and Ecology, University of California, Davis, CA 95616 volution has done wonders with lions of years (Ma) earlier than the first the extant species, which also affects the mass extinctions. Organic diver- body parts of leaf beetles are known in shape of the LTT plot. To detect signifi- sity has rebuilt itself at least five the fossil record (7). cant rate shifts, McKenna and Farrell times during the history of life applied a relatively little known test, the Eon Earth, fashioning novelty from the Diversification Rates relative cladogenesis test (9), that gener- organic remnants of each catastrophe McKenna and Farrell (5) present two alizes the earlier and more widely used (1). Modern biodiversity arose from a main conclusions regarding diversifica- Slowsinski–Guyer (SG) sister-group global extinction event at the end of the tion rates in Cephaloleia. First, they in- test (10). This test looks for unusually Cretaceous period (the ‘‘K-T’’ extinc- fer an across-clade slowdown in rate species-rich clades in the cohort of tion) that probably was caused by an over time. Second, they localize the clades all originating at one slice of time asteroid or comet impact. This penulti- most significant shifts in diversification in the tree and can be more powerful mate mass extinction [that before the rate to the Eocene, with some addi- than the SG test in identifying subtle recent and incipient human-caused ex- tional shifts in the Oligocene in nested shifts owing to the replication afforded tinctions (2)] eliminated 80% of large clades arising subsequent to one of these by the cohort beyond the two clades in plant species, 20% of marine animal shifts. These analyses are simultaneously the SG test. families, all dinosaurs except ancestors a tour de force of statistical phyloge- Naturally, the strength of this chain of present-day birds, and probably many netic techniques and a cascading set of of inferences depends on each link. specialized insect–plant associations dependencies of one inference on an- There is little reason to question the (3). Some of the lineages that pushed other. At the head is a phylogenetic tree basic phylogenetic results; they rest on through the end-Cretaceous bottleneck of the insects, relying on a now-standard now-standard technology. The diversifi- have done extremely well during the en- set of molecular sequence markers and cation rate analyses do, however, all suing 65 million years of often favorable the conventional arsenal of tree-building depend critically on the estimated diver- Cenozoic environments. Bony fishes, gence times. In any ‘‘reconstructed tree’’ corals, and several invertebrate groups (i.e., a tree in which any extinction lin- enjoyed net cladogenesis and speciation Rolled leaves of eages are removed, as in any tree built that greatly exceeded extinction. On entirely from sequence data; ref. 11), land, groups with a good start before host taxa are crucial the lineages-through-time plot will imply the end-Cretaceous extinction and im- that the ages of nodes are pulled toward pressive radiation since include birds, to the definition of the root if few terminals are sampled, mammals, insects, and angiosperms. pulled toward the recent if extinction is Stebbins (4) inspired modern research ecological niches. rampant, and pulled in either direction on recovery from extinction with the depending on whether and where spe- metaphor that plant communities are ciation or extinction rates change over simultaneously ‘‘museums,’’ which pre- techniques. Next is an analysis of diver- time. In a relative cladogenesis test, the serve ancient lineages, and ‘‘cradles,’’ gence times of the nodes in this tree by number of lineages present at any in- which foster speciation. The power using penalized likelihood, one of a stant in time is also directly determined of science to address these issues has by the divergence times of the tree’s growing set of so-called ‘‘relaxed clock’’ been greatly increased by the advent of nodes. procedures that attempt to infer clade molecular phylogenetics, which when If molecular clocks were common fea- ages from sequence data without a strict complemented by paleontology, bioge- tures of evolution, then given enough requirement of clocklike molecular ography, evolution, and ecology, can data, one could estimate all divergence rates. This analysis combined calibra- give a picture of the tempo of recovery. times to within a level of accuracy de- tions from three insect fossils, used as In this issue of PNAS, McKenna and termined basically by the error in fossil Farrell (5) address the temporal trajec- minimum age constraints within the in- calibration dates. Absent a clock, the tories of diversification rates and, by group, and a range of estimated ages problem is harder, but there has been inference, the mechanisms of recovery (120–145 Ma) of their monocot host no shortage of statistical muscle applied from the end-Cretaceous mass extinc- plants as maximum age constraints. Fi- to it. Unfortunately, theoretical results tion of the specialized and speciose nally, based on the resulting ‘‘chrono- (11) indicate, unlike in the clock case, Cephaloleia leaf beetles of Middle gram,’’ a plot of the number of lineages that the accuracy of any given diver- America. through time (LTT) was used to exam- gence time cannot be improved beyond In a previous article (6) McKenna ine the tempo of speciation and extinc- a certain point with the mere addition and Farrell brought together intensive tion across the group. The shape of this of more sequence data; instead, some taxonomy and the basic molecular phy- plot can reveal relative speciation and combination of additional fossils or ad- logeny of this genus and its relatives. extinction rates and in some cases detect ditional genes whose pattern of lineage- These are ancient beetles that have long changes in one or both. However, LTT specific rate variation is different is specialized upon a single clade of mono- analyses must also contend with the cots, the Zingiberales. The distinctive confounding interactions between spe- feeding damage of Cephaloleia ancestors ciation rates, generating the observed Conflict of interest statement: No conflicts declared. on fossilized ginger leaves dating from diversity; extinction rates, which lead to See companion article on page 10947. late in the Cretaceous and early in the a censoring of the observed diversity *To whom correspondence should be addressed. E-mail: Eocene demonstrated that these radia- and some odd pathologies in the statisti- [email protected]. tions of leaf beetles began some 20 mil- cal analysis (8); and partial sampling of © 2006 by The National Academy of Sciences of the USA www.pnas.org͞cgi͞doi͞10.1073͞pnas.0603510103 PNAS ͉ July 18, 2006 ͉ vol. 103 ͉ no. 29 ͉ 10827–10828 Downloaded by guest on September 27, 2021 necessary. These results are in accord dant recent Miocene–Pliocene diversifi- portant role in defining the number or with much empirical work on divergence cation of Cephaloleia, timed with the breadths of the niches that would pro- times that argues for more attention to closing of the oceanic gateway between tect the different species from competi- the number and quality of the fossil cali- North and South America, generated a tive exclusion for so many millions of brations (12). Unfortunately, present ‘‘crown’’ diversification of Cephaloleia years. assessments of confidence intervals on coincident with the crown diversification divergence times are not able to over- of Heliconia hosts. McKenna and Farrell Natural Enemies and Niches? come systematic biases or model mis- find that the extant sister species of McKenna and Farrell (5) posit that new specification (with respect to rate Cephaloleia arising in the Eocene (Ͻ33.9 niches were crucial to the radiations changes), which leaves us able to do Ma) in their analysis are allopatric or that they demonstrate. If Cephaloleia’s little else but explore the robustness of parapatric, which suggests that this re- parasitoids and predators have been as the analyses to perturbation. The cent speciation has happened largely by powerful throughout the Cenozoic as present analysis uses several calibration isolation associated with the closing of they are now, could not these natural points, and it would be interesting to the Isthmus of Panama. enemies have had a hand in defining gauge the effect of each by sequentially niches and affecting diversification? removing them. Niches and Diversification While ‘‘top-down’’ forces are known to From an ecological perspective, an ex- be hugely important in protecting tropi- Host Plant and Beetle Radiation tremely interesting aspect of McKenna cal plants from insect herbivory (14), McKenna and Farrell (5) infer that mul- and Farrell’s results (5) is the apparent ‘‘enemy-free space’’ (15) has been sub- tiple extant clades of Cephaloleia radi- convergence long ago to the habit of sidiary to allelochemistry, host plant ated in Paleocene–Eocene (circa 55–44 living and feeding within rolled Helioco- speciation (5, 16), and geographical sep- aration (5) in theories of speciation. Is Ma) times of global warming when their nia leaves with Cephaloleia of beetles in it possible that temporary enemy-free Zingiberales host plants radiated and two related genera, Alurnus and Chelo- space was found by Cephaloleia upon expanded their geographical ranges to basis. The authors suppose that