Abstract The spontaneous mutation rate (μ) is shaped by two distinct forces: the passage of chronological time, inevitably associated with mutation accumulation, and the speed of development, where rapid replication contributes to error rates. Disentangling these forces has been a major challenge, particularly in ectotherms. Testing the distinct predictions of the classic replication-dependent generation length hypothesis and the time-dependent accumulation model, we experimentally assessed individuals with naturally short (15 days) and long (38 days) generation time (GT), which is mainly determined by differences in developmental speed in the midge Chironomus riparius. To overcome the challenges of swarm-mating, we employed a parent-offspring pedigree reconstruction approach using pooled sequencing of siblings to infer parental allelic composition. We estimated the de novo mutation rates by whole genome sequencing. We found that Long-GT groups accumulated 1.2-fold more total mutations per generation (μ/gen), consistent with time-dependent mutagenic processes. Conversely, Short-GT groups exhibited a nearly two-fold higher mutation rate per day (μ/day) and a trend toward a transition-biased mutational spectrum (Ts/Tv ratio = 1.14 vs. 0.89), a pattern consistent with a replication-dependent errors in DNA. These results suggest that the overall mutation load is the product of these two interacting processes. Integrating our data with previous studies and life-history data, we show that the daily mutation rate followed a non-linear relationship with respect to developmental speed, and that the species’ generation time mode coincides with its minimum. This suggests that the developmental speed is, amongst other factors, selected to optimize the mutational load by balancing between replication accuracy and developmental speed.
Bulut et al. (2026) studied this question.