3 RESULTS AND DISCUSSION

Using the midpoint between initial Ts and final Ts, differences in Ts were noted between treatment groups (table 2). The Fmax-test for homogeneity of variances (Fmax = 3.807) showed no significant differences (P ≥ 0.05) among groups. The analysis of variance showed Ts differs significantly (F = 30.19) between groups. The Student-Newman-Keul procedure identified five nonsignificant subsets, with groups 1 and 4 each being significantly different from all other groups (table 2).

TABLE 2 SHRINKAGE TEMPERATURE ANALYSES (GROUPS FROM TABLE 1 LISTED IN DECREASING ORDER BY MEAN VALUE)

The most stable condition for collagen was provided by the untreated control group, which had a mean Ts of 71.3. This higher than expected value (see Haines 1987) may demon-strate a need to study Ts of skin tissues of nonconventional species.

The fact that the mean Ts of group 1 (control) is significantly greater than that in the other groups indicates that treatments involving desiccation or freezing generally contribute to collagen destabilization. A comparison of the group means shows, however, that groups desiccated first tend to be more stable than groups that were frozen first. This trend may be related to the denaturization of some proteins by phase transitions of water during freeze-thaw cycles, unfrozen water catalyzing hydrolytic or enzymatic activity, or radicals from lipid degradation (Matsumoto 1980; Florian 1990b).

Certain treatments are particularly relevant to museum practices. For instance, group 6 (dried and frozen) simulates current freezing practices for pest control (Florian 1990a). Although the mean Ts of group 6 is slightly lower than group 2 (dried only), it is reassuring to learn that the difference is not significant. These data suggest freezing dried skin tissues has minimal effect on collagen stability. Similarly, a comparison of group 2 (dried only) to group 5 (dried, rehydrated, and frozen), provides information regarding freeze-dry treatments of skin and leather that have become water-saturated, possibly as a result of disaster situations. Again, it is reassuring to find that freezing of water-saturated skin material had no significant effect on collagen stability. Although group 5 had a slightly higher mean Ts than group 2, the difference was not significant and could be explained by three Ts values of group 5 having higher than expected final temperatures.

A comparison of the Ts of groups 3 and 4 (frozen first) with the Ts of group 5 (dried, rehydrated, and frozen) is interesting because each group was frozen with unbound water present. The difference is significant. One explanation is that the phase transition of water in fresh skin involves intramolecular space, thus hydrogen bonding is negatively affected, as some unbound water may be expanded in a frozen state. Once the skin has been desiccated, it is possible that the addition of unbound water through rehydration processes involves primarily intermolecular spaces; thus the phase transition of water has little or no effect on hydrogen bonding. An alternative explanation is that water exists in the intramolecular spaces, and most is not frozen at the temperatures used in this study (Florian 1990b). If this is the case, however, collagen stability might be affected by other degradation processes.

The findings from group 4 (frozen, thawed, and dried) suggest a need for altering the common practice, particularly in vertebrate natural history collections, of freezing animals so that standard preservation techniques can be performed at a later time. The Ts for group 4 (mean = 66.1) was significantly lower than all other groups. The fact that the Ts was significantly lower than that of group 3 suggests that the freezing and subsequent desiccation had cumulative negative effects on collagen stability. Florian (1990b) also has recognized preservation problems that stem from freezing fresh skin tissues.

After visually inspecting skin or leather subjected to freezing treatment and noting little or no change, one might question why there would be a concern for a reduction in Ts. One concern might be the possible cumulative effects of treatment, especially treatment given to group 4. Perhaps the greater concern is that the loss of collagen stability potentially makes the material more susceptible to attack by proteolytic enzymes and possibly other mechanisms of deterioration (Stroz et al. 1993; Glew et al. 1994). Degraded skin material may also deteriorate faster than more stable forms. Young and Grimstad (1990) suggest that this possibility deserves further investigation. To fully appreciate the importance of the noted losses in Ts, it would be desirable to have a better understanding of the lower Ts limits provided by severely degraded collagen to evaluate the impact of treatments in terms of percent change to the potential degradation possible.

While freezing for pest control and for stabilizing water-saturated skin material may have minimal effect on collagen stability, it is important to realize that other components of the skin or leather, such as lipids, may not demonstrate the same level of stability (Florian 1990b). Furthermore, additional testing is needed to determine if repeated treatment, such as freezing for pest control, has any deleterious effects on the stability of collagenous tissues.