Atmospheric oxygen partial pressure (aPO₂) has been hypothesized to contribute to the evolution of insect size. Indeed, Palaeozoic hyperoxia correlates with insect gigantism while gigantism is absent during the subsequent Triassic hypoxia. We tested this assertion with two multigeneration oxygen rearing experiments. First Drosophila melanogaster were subjected to laboratory natural selection at 10, 21 and 40 kPa aPO₂ while sizes were monitored every generation. Over multiple generations the flies’ sizes diverged showing a positve relationship of size with aPO₂. Increases and decreases in size were mediated via combinations of developmental plasticity, parental effects and evolution. These results do suggest that aPO₂ affects size and in some cases changes in size have fitness benefits.

In the second experiment D. melanogaster were again reared in 10, 21 and 40 kPa aPO₂ but now we selected for size. The largest individuals in one generation founded the following generation. After multiple generations of size selection, 21 and 40 kPa flies showed marked but similar size increases. The 10 kPa flies, by contrast, decreased in size in the initial generations exposed to 10 kPa and only then showed notable but slight size increases. These results suggest that lower aPO₂ does indeed limit the potential maximal size in D. melanogaster.

These experiments do suggest that aPO₂ has an important effect on body size determination. Diffusive oxygen transport and the negative relationships of certain tracheal dimensions with aPO₂ might be traded off for size during development, and over multiple generations become fixed in the population.