Cope’s rule in the evolution of marine animals

See allHide authors and affiliations

Science  20 Feb 2015:
Vol. 347, Issue 6224, pp. 867-870
DOI: 10.1126/science.1260065

Getting bigger all the time

In today's world, many animal species are large, with even larger species only recently extinct, but the first animals to evolve were tiny. Was this increase in size due to active selection or to some more random process? Heim et al. test the classic hypothesis known as Cope's rule, which posits that there is selection for increasing body size. They analyzed a data set that spans over 500 million years and includes more than 17,000 marine animal species. In support of Cope's rule, body volumes have increased by over five orders of magnitude since the first animals evolved. Furthermore, modeling suggests that such a massive increase could not have emerged from a random process.

Science, this issue p. 867


Cope’s rule proposes that animal lineages evolve toward larger body size over time. To test this hypothesis across all marine animals, we compiled a data set of body sizes for 17,208 genera of marine animals spanning the past 542 million years. Mean biovolume across genera has increased by a factor of 150 since the Cambrian, whereas minimum biovolume has decreased by less than a factor of 10, and maximum biovolume has increased by more than a factor of 100,000. Neutral drift from a small initial value cannot explain this pattern. Instead, most of the size increase reflects differential diversification across classes, indicating that the pattern does not reflect a simple scaling-up of widespread and persistent selection for larger size within populations.

Body size constrains key ecological and physiological traits such as generation time, fecundity, metabolic rate, population size, and home range size (1, 2). Because of perceived advantages associated with larger size, there has long been speculation that animals tend to increase in size over evolutionary time (38), a pattern commonly referred to as Cope’s rule. Fossil data support size increase in many cases (6, 915), but numerous counterexamples also exist (1622). Moreover, some instances of size increase could simply result from neutral drift away from an initially small size rather than requiring any active selection for size (17, 22).

To determine whether animal sizes have increased since the start of the Cambrian [542 million years ago (Ma)] and, if so, whether the increase can be accounted for by neutral drift or requires active evolutionary processes, we compiled adult body size measurements for 17,208 genera of marine animals from the phyla Arthropoda, Brachiopoda, Chordata, Echinodermata, and Mollusca, with stratigraphic ranges in the fossil record resolved to stages, the finest temporal units in the global geologic time scale (23) (fig. S1A). These phyla together account for 74% of animal diversity in the fossil record (24), and our data set covers 75% of total known genus diversity in these phyla. We measured the three major body axes from published images of specimens (typically holotypes of the type species) (23) in order to estimate the size of each genus as a simple geometric solid or from known length:mass relationships (23). We used linear regressions of biovolume on maximum length for classes and phyla to estimate the biovolume of genera for which fewer than three major axes were illustrated (23) (fig. S2 and table S1).

Figure 1 illustrates the sizes of the marine animal genera in our data set across the past 542 million years. The mean biovolume across genera has increased by more than a factor of 150 (2.18 log10 units; the median increased by 2.35 log10 units) since the earliest Cambrian. Over the same interval, the range in biovolume expanded from 8 orders of magnitude in the Cambrian to 14 orders of magnitude in the Pleistocene (1 Ma). Most of this expansion in size range reflects an increase in the maximum, which climbed by more than three orders of magnitude between the Early Cambrian and Middle Devonian (542 to 385 Ma) and by an additional two orders of magnitude thereafter. In contrast, the overall minimum size decreased by less than one order of magnitude between the Early Cambrian and Middle Devonian and has remained stable ever since.

Fig. 1 Body size evolution across the past 542 million years.

The distribution of fossil marine animal biovolumes across the Phanerozoic is shown. The colored horizontal lines show genus durations. The thick black line indicates the stage-level mean body size. The thin black lines demarcate the 5th and 95th percentiles. Cm, Cambrian; O, Ordovician; S, Silurian; D, Devonian; C, Carboniferous; P, Permian; Tr, Triassic; J, Jurassic; K, Cretaceous; Pg, Paleogene; N, Neogene.

To test models of neutral change relative to active processes, we compared observed trends in maximum, mean, and minimum size to expectations generated by three evolutionary branching models: an unbiased random walk (i.e., Brownian motion; fig. S3A), a bounded random walk (i.e., Brownian motion with a reflecting lower bound; fig. S3B), and a size-biased random walk (fig. S3C) (23). The size-biased model fit observed trends in the minimum, mean, and maximum size better than the neutral and lower-bounded models (Fig. 2 and fig. S4). The observed minimum size is within the predicted range of all three models, but the observed mean and maximum sizes trended above the predicted range for the unbiased and lower-bounded models (Fig. 2 and fig. S4). As an additional test of the three models, we compared the observed distribution of biovolumes of all 2280 Pleistocene genera, the most recent and taxonomically diverse time interval in our analyses, to distributions predicted by our branching models and found strong support for the size-biased model over the other two models (table S2) (23). Finally, we used an independent, likelihood-based approach to compare support among five models for the trend in mean size over the entire 542 million years (23). Three models assumed a single mode of size evolution across the Phanerozoic (driven trend, random walk, or stasis). The best-supported of these models was the driven trend toward larger size (Table 1). However, allowing a shift in model type and parameters at the era-bounding mass extinctions improved the overall model fit, with the best-fit model in the Paleozoic being a driven trend toward larger size, followed by stasis in the Mesozoic, and a reversion to a driven trend in the Cenozoic (23) (Table 1). Removing the marine tetrapods, which tend to be very large, did not change this result (table S3). Thus, most of Phanerozoic time has been characterized by a trend toward larger animal sizes.

Fig. 2

Comparison of observed biovolume trends to those obtained from stochastic branching models (23). The colored regions highlight the size space occupied by 90% of the 1000 model runs. For clarity, only results for the size-biased (red) and unbiased (blue) models are shown; except below the minimum size, the lower-bounded model produced results nearly identical to the unbiased model (fig. S4). (A) Minimum, (B) mean, and (C) maximum sizes. Time scale abbreviations are the same as in Fig. 1.

Table 1

Results of model comparisons for the Phanerozoic trend in mean biovolume. Lower corrected Akaike information criterion (AICc) and higher Akaike weights indicate more support for a given model. logL is the log likelihood, and K is the number of free parameters in each model. The two-phase model has a break point at the Permian/Triassic boundary. The three-phase model has break points at the Permian/Triassic and Cretaceous/Paleogene boundaries. The best-fit model for each phase is used in multiphase models. The three-phase model is the best-supported model. See the supporting materials for details of the statistical methods (23). n/a, not applicable.

View this table:

The trends in minimum, mean, and maximum biovolume of marine animals are consistent with actively driven size increase and not consistent with simple neutral drift away from an initially small ancestor. To determine the extent to which this trend reflects size increase at low taxonomic levels across all phyla versus differential diversification of higher taxa with different mean sizes, we compared the observed trend in the mean to the expected trend if size were kept constant as diversity changed within different levels of the Linnaean hierarchy (fig. S5). This comparison demonstrates that much of the observed size increase reflects differential diversification among classes and is consistent with hierarchical size evolution in Paleozoic brachiopods (15). This finding also sheds light on the tripartite nature of the Phanerozoic trend in mean biovolume, as there was little differential diversification among classes during Mesozoic time (fig. S1B).

The dominance of differential diversification at the class level in producing the overall trend toward larger animal sizes emphasizes the hierarchical nature of evolutionary processes. In addition, it suggests that the widespread bias toward selection for larger size observed in extant populations (8) is unlikely to propagate into large-scale evolutionary trends observed in the fossil record. If the Phanerozoic trend reflected widespread selection for larger size at low taxonomic levels, we would expect to see most of the size trend explained by size increase within families, the lowest taxonomic level at which we can aggregate our data. These findings suggest that the factors favoring the overall trend toward larger size in marine animals relate to basic body plan and ecological life mode rather than competitive advantages associated with size differences within populations. These findings do not rule out an additional, smaller component related to widespread selection for larger size within populations, as there is also a component of size increase that occurs within families (fig. S5).

The taxonomic composition of the smallest and largest genera over time suggests the operation of constraints at the size extremes, even if these constraints are not required to model the overall distribution of animal sizes. The size minimum has been populated nearly exclusively by ostracods (a class of exclusively small-bodied crustaceans) since Silurian time (420 Ma). With the exception of the Middle/Late Triassic (235 Ma), where there is a maximum in ostracod size and a minimum in other animals, no other group in our data set comes within a factor of 6 of the smallest ostracod. The size maximum has been populated entirely by chordates since the Early Triassic (252 Ma), with no other genus since then coming within a factor of 2.5 of the largest chordate.

The dominance of a single phylum at each end of the size spectrum could result from simple incumbency effects, but transitions over time in the class affinities of the largest marine chordates suggest that physiology is also an important constraint, at least on the overall maximum size of marine animals. Nearly all of the largest solitary marine bilaterian genera have been reptiles and mammals. Tetrapods first reinvaded the oceans during Late Permian time (260 Ma) and rapidly occupied the size maximum (Fig. 3). Reptiles continued to dominate the top end of the size spectrum during the Mesozoic. Cetaceans were the first mammals to evolve a marine lifestyle and have occupied the largest marine body sizes since they first invaded the oceans during the Eocene (48 Ma) (Fig. 3). Air breathing is an exaptation (25) that can explain the rapid and widespread attainment of large size in marine reptiles and mammals. Relative to water, air has 20 to 30 times the concentration of O2, is up to 100 times less viscous, has diffusion rates of O2 through membranes that are 300,000 times faster, and is about 1000 times less dense (26). Thus, large animals are better able to meet their metabolic needs by breathing air than by breathing water. In fact, O2 limitation has been proposed as a mechanism for limiting the evolutionary emergence of large, free-swimming, predatory bilateria generally (27, 28).

Fig. 3 Taxonomic compositions of the largest genera.

Fishes and, later, air-breathing tetrapods, dominate the top of the size distribution. All genera with epoch- or stage-resolved stratigraphic ranges are plotted here, allowing for the inclusion of more large vertebrates. Horizontal lines show genus durations. The heavy black line demarcates the 95th percentile of all genera. Time scale abbreviations are the same as in Fig. 1.

Synoptic size data show that the average size of marine animals has increased substantially since the Cambrian and that this increase reflects differential diversification of large-bodied classes rather than neutral drift. A remaining question is the extent to which this differential diversification was enabled by intrinsic factors such as physiology, escalatory interactions between predators and prey (29), or changes in the physical and non-animal environment, such as oxygen availability (30) or the amount and quality of primary production (7). Testing among these controls will be critical to understanding how the physical and biological environments combine to shape the evolution of global ecosystems.

Supplementary Materials

Materials and Methods

Figs. S1 to S6

Tables S1 to S3

Caption for Database S1

References (31120)

References and Notes

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
  85. 85.
  86. 86.
  87. 87.
  88. 88.
  89. 89.
  90. 90.
  91. 91.
  92. 92.
  93. 93.
  94. 94.
  95. 95.
  96. 96.
  97. 97.
  98. 98.
  99. 99.
  100. 100.
  101. 101.
  102. 102.
  103. 103.
  104. 104.
  105. 105.
  106. 106.
  107. 107.
  108. 108.
  109. 109.
  110. 110.
  111. 111.
  112. 112.
  113. 114.
  114. 115.
  115. 116.
  116. 117.
  117. 118.
  118. 119.
  119. 120.
  120. Acknowledgments: We thank G. Griggs, M. Faerber, M. Laws, S. Sanghvi, L. Taylor, and the many undergraduate and high-school students for making body size measurements. J. Saltzman helped recruit high-school students. G. Hunt assisted with time series analysis. A. Clauset and M. A. Etnier kindly made their body size data on extant marine mammals available. Funding was provided by NSF grant EAR-1151022, the Stanford School of Earth Sciences, and the Swarthmore College James Michener Faculty Fellowship. Raw data files used for all analyses are permanently archived in the Stanford Digital Repository ( This is Paleobiology Database publication 217.
View Abstract

Navigate This Article