Experimental studies on supercooling in two Antarctic micro-arthropods

Samples of mites and Collembola which had been acclimated at 5°C and provided with natural foods were cooled at four constant cooling rates: 1, $\text{1}\text{2}$, $\text{1}\text{4}$, $\text{1}\text{8}\text{deg min}{}^{\mathrm{?1}}$ and ca 20 deg min^?1, and their individual supercooling points measured. Frequency distributions of supercooling points comprised not less than 84 (Alaskozetes antarcticus) and 96 (Cryptopygus antarcticus) individuals in each case. Two modal groups were displayed in these distributions, which were widely separated in temperature and termed low group and high group. In Alaskozetes a trough between ?3 and ?4°C was present in the high-group distribution, which may be due to a lack of a certain class of nucleators. The highest temperatures at which animals froze occurred at the slowest cooling rate ($\text{1}\text{8}\text{deg min}{}^{\mathrm{?1}}$), whereas rapid cooling removed the trough to form a single high-group peak. In Cryptopygus, the high groups were narrow and peaked (<2 deg wide) at all cooling rates, with a downward shift of ca 1 deg between the rates $\text{1}\text{8}$ and $\text{1}\text{2}\text{deg min}{}^{\mathrm{?1}}$. Both species showed a trend towards a lower mean low-group supercooling point at faster rates of cooling, but these were not significant. Regressions of cooling rate on individual low-group supercooling points (≥?20°C) for both species showed a significant negative correlation, which did not differ between species. The distribution of the deviations about each rate-defined mean in the low group for each species was skewed to the right, with 88% occurring between ±2 deg of the means. It is suggested that minor deviations (e.g. halving or doubling of the cooling rate) do not affect the resultant supercooling points at non-constant cooling rates, but a rate of 1 deg min^?1 is to be preferred.