During the late summers of 1997, 1998, and 1999, freshly laid egg masses were placed into individual nylon screen pouches and attached in grid arrays to the trunks of two large oak trees (60cm or larger diameter) at two experiment sites: near Augusta, MI, in the southwestern Lower Peninsula of the state, and near Roscommon, MI, in the north central Lower Peninsula. Egg masses were placed on each of the trees at each of four cardinal direction at four heights: 0m, 1m, 2m, and 3m. Egg mass temperatures and other meteorological data were monitored each season continuously through the following winter season. In the spring of each season, a portion of the egg masses from each height and aspect combination were collected at 3 wk intervals and returned to the lab to be reared to eclosion or examined to determine cause of death.
During the three seasons of the experiment, both height and aspect were found to strongly influence temperature variability of the egg masses, largely as a result of differential rates of solar loading. Egg masses on the southern side of trees experienced the highest maximum temperatures, the greatest deviation from ambient air temperature, and the greatest variability in diurnal temperatures, followed by egg masses on western and eastern aspects. Highest maxima occurred in egg masses on the southern and western aspects, especially near the buttresses of the trees. Minimum temperatures were significantly warmer when egg masses were covered by snow, but snow cover also appeared to damp out maximum temperatures, especially on the southern side of trees. Overall, warm fall temperatures had a greater effect on egg survival than winter temperatures. Egg masses on southern exposures had the highest egg mortality, while egg mortality was consistently lowest on the northern exposures. Egg survival on the eastern aspect was similar to the northern aspect, and survival on the western aspect was midway between northern and southern aspects. Egg masses near the ground sustained greater mortality than egg masses at 1,2 or 3 m on all but the northern aspect, where height had little effect.