Although attempting to consider the impact of freeze-thaw, the monitoring of associated thermal conditions must, effectively, be context free. This is important for several reasons. First, although freeze-thaw is being evaluated, data must be of a nature that also allows determination of the spatial and temporal role of other processes. Second, by being ‘holistic’ rather than (assumed) ‘specific’ in character the data do not pre-determine the outcomes. Third, by being able to be used for evaluation of multiple processes the data are, paradoxically, of a nature that facilitates a more detailed understanding of freeze-thaw activity itself. In considering freeze-thaw it is important to recognize that the process is not a singularity but rather comprises a range of mechanisms, each determined by an interaction between thermal and moisture conditions with the properties of any given building material. In the absence of (the required) moisture data, the thermal data need to be adequate to validate, or invalidate, specific mechanisms, as well as to offer indirect proxy information indicating whether or not some form of freeze-thaw weathering indeed took place. Significant in this regard are not just freeze-thaw amplitudes and durations but also the associated rate of change of temperature (⊿T/⊿t). In addition, the record rate must be fast enough to monitor (should they occur) leased as water turns to ice; this being a proxy identification that water was present and did indeed freeze (the finding of ‘zero a proxy for the existence of water that froze within the material). In reality, the thermal data acquisition requirements for monitoring of both ⊿T/⊿t rates and exotherms are effectively the same; namely high-frequency thermal monitoring at an interval of at least one-minute. Thermal monitoring at one-minute intervals may have produced logistical problems in the past but modern data loggers with multiple channels, long-life battery power and large storage capacities can handle such requirements with ease. The resulting data are of a temporal nature that allows for the evaluation of thermal stresses, especially those associated with ⊿T/⊿t events ≥2 ℃min^-1. Equally, such data are able to resolve the short-interval heat transfer associated with latent heat release at phase transfer. Significantly, such data not only identify the occurrence of exo-therms but also show the temperature at which freezing took place. Further, given sub-zero rock temperatures, the absence of exotherms shows when thermal conditions may have been suitable but no water was available to freeze or despite water being present it did not freeze. Thus, this approach provides objective data allowing for the true counting of actual events rather than the subjective counting based on the assumptions that (a) water was indeed present and (b) that it froze within a certain thermal range. This latter approach (assumed counting) has now been shown to suffer from potentially massive error, especially within a spatial context. In respect of thermal monitoring, the key prime requirements are: large data capacity loggers with multiple channels, high resolution loggers, high-frequency logging capacity, high resolution transducers, fast response time transducers, large spatial distribution of transducers (including, where possible, with depth within the material being monitored). In terms of transducers, experiments have suggested that 40 gauge thermocouples satisfy resolution (0.1℃) the almost invisible nature of the wire, not impacting on the aesthetics of a site. If drilling of the building material (for the emplacement of transducers) is possible, holes are less than 0.2 mm in diameter, or, if a predrilled block is situated at the site, the visual impact is still very small. Where any form of attachment or drilling is prohibited, infra-red (IR) sensors are now of a resolution (0.1℃) response time (0.002 sec) and monitored area (1 mm²) that can provide excellent data; but at an aesthetic cost during monitoring. The use of infrared sensors is ideal for monitoring of surface pigments (e.g. in cave art) or fragile components unsuitable for direct contact sensors. Once collected, data have shown that many pre-conceived notions, especially in respect of freeze-thaw, are in error. Despite cold temperatures (the common “indicator” for the assumed occurrence of freeze-thaw) data have shown that either water was not present in the rock to freeze or it simply did not freeze at the available temperature; equally the temperature at which freezing occurs has been found to often be colder than the assumed value. Sometimes the freezing of water was found to be progressive with depth while at other times it was instantaneous over the outer several centimetres of the rock. Spatial and temporal variability of freeze-thaw events were both extremely large. Thermal stress events often, in magnitude, frequency and spatial distribution, exceed freeze-thaw in terms of number of occurrences. Further, moisture and thermal conditions show that, in cold environments, chemical weathering can occur for long periods – perhaps all winter. Finally, as much as these data help us to go forward in our understanding of weathering, they still need to be directly linked to actual breakdown – we cannot simply assume that because freeze-thaw may occur it is (in the absence of proof) the cause of the damage we observe. This is the next step – the connectivity of material failure with specific process.