ON THE ASSOCIATION BETWEEN MACROECOLOGY AND MACROEVOLUTION

Although it is widely believed that the ecological interactions are fundamental determinants of macroevolutionary trends over the Phanerozoic, such ideas have rarely been tested because appropriate data sets are scarce. Furthermore, standard measures of diversity should be complemented with metrics for abundance, because these measures are often decoupled and can lead to different conclusions. Using the Paleobiology Database, we were able to examine a range of potential macroecological/macroevolutionary hypotheses including the "escalation hypothesis" and the "bulldozer hypothesis." We scored each marine invertebrate taxon by three ecological categories (life habit, diet and locomotion) and divided the Phanerozoic into 54 ten-m.y. time-bins each averaging 3472 taxonomic occurrences (our surrogate for abundance) and 240 taxa. To avoid the problems associated with sampling intensity bias that plague estimates of global diversity, we used the proportion of taxa or occurrences with each ecological score within each time bin. We found strong correlations between first order differences of consecutive counts of occurrences and taxa for each ecological grouping, so coupling of hierarchical levels is apparent. Secondly, a simple coupled model of independent exponential increase or decrease fitted well to changes in the proportions of taxa and occurrences of the various ecological groupings over the Phanerozoic; model trajectories strongly supported well-known macroevolutionary hypotheses. However, using differenced data, cross-correlations between ecological groupings illustrated weak or nonexistent associations in all cases of macroevolutionary interest, implying that ecological interactions are not an important driver of the macroevolutionary trends. Strong associations only exist between implicitly dependent groups (e.g., carnivores are mobile; suspension feeders are stationary). To check for time-lags in evolutionary responses to ecological forcings, correlations were calculated for several lag periods in either direction and illustrated that associations were similarly weak or weaker. Thus, the ecological groupings we defined may indeed be ships passing in the night.