Geographic variation in embryonic development time and stage of diapause in a grasshopper.

Embryonic development times and the stage at which embryonic diapause occurs varied dramatically among 23 populations of the Melanoplus sanguinipes/ devastator species complex in California, USA. Grasshoppers were collected from a wide range of latitudes (32°57N to 41°20N) and altitudes (10m to 3031 m), spanning much of the variation in climatic conditions experienced by these insects in California. When reared in a "common garden" in the laboratory, total embryonic development times were positively correlated to the mean annual temperature of the habitat from which the grasshoppers were collected (varying from about 19 days to 32 days when reared at 27°C). These grasshoppers overwinter as diapausing eggs and the proportion of embryonic development completed prior to diapause was significantly higher in populations collected from cool habitats (>70%) than in populations collected from warm environments (<26%). The length of pre-diapause development time is determined by the stage of embryonic development at which diapause occurs, and varies considerably among populations of these grasshoppers; grasshoppers from warmer environments tend to diapause at very early stages of embryogenesis, while grasshoppers from cooler environments diapause at very late stages. The combined effect of variation in embryonic development times and variation in the stage at which diapause occurs results in a dramatic reduction in the time needed to hatch in the spring; populations from warm environments required up to 20 days (at 27°C) to hatch while populations from cool environments required as few as 5 days to complete embryonic development prior to hatching. Egg size also varied significantly among populations, but tended to be larger in populations with shorter embryonic development times. Significant family effects were observed for development time and stage of diapause, suggesting significant heritabilities for these traits, although maternal effects may also contribute to family level variation. We interpret these findings to support the hypothesis that embryonic development time and the stage of embryonic diapause have evolved as adaptations to prevailing season lengths in the study populations.