A NEW METHOD FOR RECONSTRUCTING PATTERNS OF INTRA-ANNUAL GROWTH IN BIVALVE MOLLUSKS

Bivalve mollusk shells contain archives of important biological and environmental data. For example, widths of periodic growth increments record growth rates, and stable oxygen-isotope (δ¹⁸O) values vary as a function of temperature. Reconstruction of temperature dependent intra-annual growth rates can be accomplished by counting and measuring tidal growth increments. Unfortunately, these increments are seldom preserved, making this approach difficult in fossil specimens. Here, we present a new analytical method that relates linear growth and time to reconstruct intra-annual growth rates, an important biological variable.

The basic idea is simple; the derivative of a function relating linear growth (cumulative distance) to time represents relative intra-annual growth rates – the growth function. To evaluate this idea, we simulated clam growth with different hypothetical growth functions. We used an annual temperature model based on temperature variation in the Gulf of California. Each day within the year was assigned an increment width and a δ¹⁸O value. These shells were then “sampled” assuming a 300 μm drill bit. Next, the δ¹⁸O value of each sample was converted to a temperature and assigned a date based on the temperature model. A line was then fitted to the plot of cumulative distance versus date. Finally, we graphed the derivative of the best-fit model to reconstruct the hypothetical growth function. In all cases, we accurately reconstructed the hypothetical growth function.

To further evaluate this approach, we reconstructed the intra-annual growth function of real clams (Chione cortezi) with known intra-annual growth functions. Our modeled functions reconstructed significant intra-annual growth rate variation. However, these functions did not always match the observed patterns of intra-annual growth. We believe this disagreement stems primarily from error assigning calendar dates to δ¹⁸O samples associated with high-frequency environmental variation. Our ongoing research suggests that statistical re-sampling techniques may resolve this discrepancy. This model may provide a new tool for reconstructing an important paleobiological variable and understanding ancient environmental and/or evolutionary patterns.