The new compound [Cu(cyclam)(H 2 O)]-{[Cu(cyclam)] 2 [HTiNb 9 O 28 ]}•26H 2 O (1) (cyclam = 1,4,8,11-tetraazacyclotetradecane) was obtained under solvothermal conditions. Its crystal structure contains a monotitano-nonaniobate anion in which one position is equally occupied by Nb(V) and Ti(IV). The anions are expanded by [Cu(cyclam)] 2+ cations via Nb−O−Cu bridges generating {[Cu(cyclam)] 2 [HTiNb 9 O 28 ]} 2− cluster units, which are arranged into layers. Between these layers there are additionally isolated [Cu(cyclam)(H 2 O) 2 ] 2+ cations as well as hydrate water molecules. Storage of 1 at room temperature leads to loss of ∼13 water molecules, and a new crystalline phase (2) crystallizes that, with heating, transforms into the anhydrate. The reversibility of this reaction was investigated by thermogravimetry and X-ray powder diffraction (XRPD). Temperaturedependent in situ synchrotron XRPD investigations prove an abrupt phase transition, in which especially the a axis is dramatically shortened and the {[Cu(cyclam)] 2 [HTiNb 9 O 28 ] 2− } cluster is rearranged. Single-crystal X-ray diffraction of 2 reveals that, despite the unusual large shrinking of the unit cell volume, the domains formed by water removal exhibit some preferred orientation close to that expected for a topotactic reaction, which allowed the performance of a structure analysis. In the structure of 2, the two water molecules of the isolated [Cu(cyclam)(H 2 O) 2 ] 2+ cation in 1 are replaced by two terminal cluster O atoms, leading to the formation of chains via Nb−O−Cu bonds, and this phase transition is accompanied by an ordering of one of the two cyclam ligands. 50 the formation of domains, without changing the crystal 51 morphology. In most cases there is no structural relationship 52 between the hydrate and the intermediate phase, and therefore, 53 the domains show a random orientation, and the diffraction 54 pattern corresponds to that of a powder. However, in very few 55 cases there is a strong relationship between the structures of 56 the reactant and the product, and in this case the domains 57 formed in a reaction exhibit a strong preferred orientation 58 simulating the diffraction pattern of a single crystal. The 59 presence of such a reaction, called topotactic, can only be 60 proven by single-crystal X-ray diffraction, and in ideal cases 61 differences between the diffraction pattern of the pristine and 62 the product phase are only detected in the mosaic spread. 12−15 63 For such reactions the question arises what will happen if 64 larger structural changes occur for which the orientation of the 65 domains are far from perfect and if structural information can 66 be retrieved in such cases. In the course of our ongoing