Dinosaurs have recently been found in South Australia and North Alaska. At the time these places were nearer to the poles than today. Assuming that the obliquity of the Earth did not change, these dinosaurs must have lived throughout six-month-long nights. The local average annual temperature at the time has been very roughly evaluated from 0°C to 13°C (using hints from the plants and the oxygen isotopes), which is higher than today, but during winter the minimal temperature could have been around -11°C.
However, some primitive reptiles (champsosaurs) have been found at quite high latitudes from the middle Cretaceous of Canada.
The Australian polar dinosaurs were smaller than their tropical relatives, which indicates that there were important ecological constraints, one more hint in favour of a low temperature. Some of them have been shown to have had unusually big optic lobes in their brains, probably an adaptation to the polar nights.
It has been supposed that these dinosaurs did migrate to reach warmer places. In the case of Australia, however, an arm of the sea prevented them from going directly North, and they would had to run thousands of kilometres, which is the less probable as they were (for dinosaurs) relatively small. Moreover, some young remains have been found (in Alaska), making migration even less probable. They could also have hibernated, but they were probably too big to bury themselves, and (?) had no protecting fur. In any case, the developed sense of sight would have been useless if they had been away or sleeping during the polar night.Thus these dinosaurs probably remained active in the cold. Mass homoiothermy could be the solution, but (besides general considerations against it, see below) the polar dinosaurs would then have had a tendency to grow, not to become smaller than their ancestors