Fourth, we agree that oxygen is now a less important control LETTER on maximum size, at least for insects in natural settings, because of other factors (including predation). However, the claim that Reply to Dorrington: Oxygen tracheal surface area would permit an order of magnitude size concentration and predator escape increase is inconsistent with experimental results (3). Tracheal volume scales hypermetrically, not linearly, with body size (3), abilities are important controls on and constraints are imposed by specific body parts before whole- insect size body scaling limits are reached (4). Fifth, although larger Anisoptera may be favored in intra- Dorrington (1) argues that oxygen is not an important control on specific competition, Dorrington (1) notes that good maneu- insect size and that selective pressure for maneuverability and verability and higher speed play a role; thus, the relative predator escape is not inversely related to body size. Because importance of maneuverability in this instance is not clear. Most these claims are contradicted by many experimental studies, we predator evasion instead involves liftoff from a surface, and, stand by our hypothesis (2) that controls on maximum insect because smaller insects have intrinsically higher acceleration (5), size shifted from environmental (primarily oxygen concentra- they will more easily be able to escape those attacks. Greater tion) to biotic (primarily predation) following the evolution of acceleration of smaller insects will also lead to greater aerial flying predators like birds. maneuverability (although aerodynamic factors are also First, Clathrotitan is actually not unusually large for the important). modeled oxygen concentration in the Middle Triassic; in fact, Finally, comparison of living and fossil insect faunas is not it is Middle-Late Permian insects that are anomalously small. valid because of inherent differences in the method of preserva- The correlation between maximum size and oxygen will not be tion and sampling effort. Because fossil species are a small perfect because of the vagaries of insect preservation and subset of the living fauna, the fossil record will always under- sampling, and because of slight differences in the absolute ages estimate the largest taxon extant during any interval. We dem- in the time scales used (we use the current Anisian boundaries onstrated (ref. 2, SI Text) that the observed overall trend was of 247.2–242 Ma, whereas GEOCARBSULF used a previous not driven by variations in sampling of fossil assemblages, so the time scale). conclusions remain valid. Second, changes in density likely would also have affected the Matthew E. Clapham1 and Jered A. Karr size of flying insects, but would have exactly paralleled and am- Department of Earth and Planetary Sciences, University of California, plified oxygen effects. Most workers have focused on the role of Santa Cruz, CA 95064 oxygen, however, because late Paleozoic gigantism is also observed 1. Dorrington GE (2012) On flying insect size and Phanerozoic atmospheric oxygen. Proc in terrestrial arthropods and because there is abundant experi- Natl Acad Sci USA 109:E3393. mental evidence indicating that tracheal scaling is a key constraint 2. Clapham ME, Karr JA (2012) Environmental and biotic controls on the evolutionary history of insect body size. Proc Natl Acad Sci USA 109(27):10927–10930. on insect size (ref. 3 and references therein). 3. Harrison JF, Kaiser A, VandenBrooks JM (2010) Atmospheric oxygen level and the Third, although Meganisoptera may have been weaker fliers, evolution of insect body size. Proc Biol Sci 277(1690):1937–1946. 4. Kaiser A, et al. (2007) Increase in tracheal investment with beetle size supports gigantism is observed in a variety of groups. For example, hypothesis of oxygen limitation on insect gigantism. Proc Natl Acad Sci USA 104(32): the Carboniferous stem-group mayfly Bojophlebia had a forewing 13198–13203. 5. Dudley R (2000) The Biomechanics of Insect Flight: Form, Function, Evolution (Princeton length of 215 mm. Similarly, Triassic Titanoptera were much Univ Press, Princeton). larger than younger orthopteran relatives. Post-Jurassic size decreases are observed even among Anisoptera (the Late Author contributions: M.E.C. and J.A.K. designed research; M.E.C. and J.A.K. performed Jurassic Cymatophlebia suevica and Aeschnogomphus kuempeli research; M.E.C. and J.A.K. analyzed data; and M.E.C. wrote the paper. had forewing lengths of 110 mm and 106 mm, respectively) The authors declare no conflict of interest. and Orthoptera. 1To whom correspondence should be addressed. E-mail: [email protected].

E3394 | PNAS | December 11, 2012 | vol. 109 | no. 50 www.pnas.org/cgi/doi/10.1073/pnas.1215989109 Downloaded by guest on October 1, 2021