In the class of homeotypes of Si the model of electron pair bonds explains some phenomena such as the stability of Grimm-Sommerfeld phases. Similar explanations follow also from the model of electron spatial correlations. However, there are various phenomena, not explained so far by the pair model, but well interpreted by the correlations model, for instance stacking homeotypes of ZnS.r such as A1N. The correlations model has a more extended range because it considers more parameters than the pair model. A bonding type of the correlations model does not only classify the valence electron state but also properties of the core electron states. The bonding types based on the correlations model frequently lead to interesting experimental questions. Brought to you by | provisional account Unauthenticated Download Date | 6/26/15 12:56 PM Brought to you by | provisional account Unauthenticated Download Date | 6/26/15 12:56 PM Some homeotypes of Si 163promoted in the pair model because this model considers in first line the spin compensation and not the spatial correlation. The spins are already compensated in the lower bands and therefore these bands were of less interest. In the pair model the bond was something that tied the atoms together while in the correlations model the bond must be considered as a soft solid angle where global tensions cause an increased approach of the atoms. The short distances are generated by a global bonding (the correlation) because there are soft angles.Less serious differences in the nomenclature of the models may be bridged by the following correspondences.Normal adamantine structures (Pamplin, 1960;Goryunova, 1963): Stacking homeotypes of ZnS.r, briefly ShtpZnS.r. The set of S-homeotypisms should contain the 1-homeotypism (identical homeotypism) to ensure that the isotypes of ZnS.r are contained in ShtpZnS.r.Defect adamantine structures: L'ShtpZnS.r. The prime reminds to the conservation of N(, a .Normal tetrahedral structures (Goryunova 1963, Parthe 1964): R'ShtpSi.Defect tetrahedral structures: R'L'ShtpSi. Normal valence phases (Mooser, Pearson, 1959): Phases with N[ Xa = 8, An = anion, Lewis phases. The essential feature of Lewis phases is seen in their electron spin compensation, i.e. in the phenomenon that a +spin is surrounded mostly by -spins, but non-Lewis phases may have also spin compensation. Consequently, the name "normal valence phase" is not quite fortunate, the non Lewis phases are just as normal as the Lewis-phases.Polycationic phases (Pearson, 1964): Phases with Ni Aa >8, over-completed phases. There must be non-Lewis spin compensation, i.e. spin compensation between cations, and this causes smaller distances between cations, like as in the 8-Nl M = MmLn dist. rule of Hume-Rothery. The acceptable concept "polycationic phase" should not be interpreted by electron pair bonds, because also here these bonds imply a local spatial correlation, possibly contradicting to the overall spatial correlation. By the way, the concept is not appropriate for instance for the mixture Li...