A comment on this post reminded me of an occasional problem with managing environmental samples: getting them too cold.
I have two "war stories" about freezing samples:
1. I was working on an island, and a blizzard blew in as we were wrapping up groundwater sampling for volatile organic compounds (VOCs) - small vials, no headspace allowed. We ended up fleeing and leaving the coolers where they were. We came back two days later, fought our way through epic snow drifts, and found that all our VOC samples had frozen. We spent all day running around like crazy people to re-sample, and by the time we had schlepped our new batch of samples back to the dock (back through those piles of snow again!), we were so overheated that we rode in the front of the launch (about 15 degrees F on the water) the whole ride back to the mainland.
2. We had a huge batch of samples that were going to go out the next day. It was winter, but we were working out of a storage box/office combo, so we had some residual heat for the whole trailer. Our VOC samples were tucked in tight in the office, but in this case the giant 4-liter glass bottles that I so dislike broke because they were out in the storage part and too far from the heater. They were essentially ok, except for the necks. Luckily (since these samples represented probably a combined 100 hours of effort) the bottles were already wrapped in giant ziplock bags for shipping and we just pumped the water out of the broken bottles and into fresh ones.
The best weather for samples is early spring/late fall in my neck of the woods: 60 degrees F during the day for comfortable fieldwork, 35 degrees F at night to keep the samples cold, but not too cold.