The relationship between number of striae of Retzius and their periodicity in imbricational enamel formation

Imbricational crown formation times (ICFTs) estimated from the number of perikymata on tooth surfaces are error-prone because the number of days between adjacent perikymata varies across individuals and species, and is only visible within tooth microstructure. We investigated striae of Retzius (SR) numbers (analogous to perikymata numbers), SR periodicities (days between SR or perikymata), and lCFTs for a mandibular canine sample (n = 49) from medieval Denmark. We tested the relationship between SR number and periodicity to determine whether regression formulae could be produced that would allow periodicity (and ICFTs) to be determined from surface perikymata numbers. Periodicities (range = 7-11 days, mode = 8) and SR numbers (range = 142-257, mean = 190.3, s.d. = 27.5) were normally distributed; ICFTs were non-normal (mean = 1,594 days, s.d. = 65.7). We tested periodicity as a quadratic, linear, and log-log transform linear function of SR number and found an inverse relationship (quadratic: R-2 = 0.9504; linear: R-2 = 0.9138; log-log transform: R-2 = 0.9418; all p < 0.001) that allowed estimation of periodicity from SR or perikymata numbers in this population and tooth type. If periodicity and SR number are inversely related in other hominin taxa, Studies that have estimated ICFT by multiplying perikymata number by a human modal periodicity value or made inferences about development based only on perikymata numbers may have introduced substantial error into their ICFT estimates and life history inferences. The inverse relationship is similar to that predicted by a model of SR formation in which the ICFT for a given tooth type and population is held constant and all combinations of periodicity and SR number result in the same ICFT. However, we found that lower periodicities had longer ICFTs and higher periodicities had shorter ICFTs than the model predicted, suggesting that the model may not reflect the real process, or that there are other factors (e.g., sample size, misclassification, sexual dimorphism) also affecting the relationship between periodicity and SR number. (c) 2005 Elsevier Ltd. All rights reserved.