The lack of reproducibility of the cold fusion experiments, aggravated
by the great diversity and inconsistency of the positive results, imp
lies that these nuclear phenomena are hypersensitive, i.e., correlated
to a ''chaotic'' factor. All the factors considered so far, such as s
tructure, transformations, or defects of the crystal lattice; bubbles
of deuterium; dendrites, etc., are insufficiently chaotic to explain t
he known facts. Experimental data suggest that nuclear reactions take
place in active sites on the surface of the lattice, that they are sti
mulated by dynamics factors, and that they represent an extreme form o
f heterogeneous catalysis. Consequently, according to modern ideas con
cerning catalysis, the desired chaotic factor is the surface dynamics
of some metallic deuterides (hydrides). This hypothesis, called the su
rfdyn concept, is compatible with all published data, explains the pec
uliarities of cold fusion, and must be supported by an adequate theory
describing the nature and mechanisms of the different nuclear process
es.