There is much in what Ken Seddon says¹ that is beyond dispute. I mean, who can contest his statement that scientists should avoid ambiguity, and should be as clear as possible in their expositions? And since it would be mischievous to refute oneself, let me accept most whole-heartedly our 1999 definition² of polymorphism as "a phenomenon wherein the same substance exhibits different crystal packing arrangements"!

Coming to the matter at hand, the term pseudopolymorph was first suggested by McCrone,³ and it has entered the pharmaceutical literature in a big way. Both Byrn⁴ and Bernstein⁵ have mentioned the term in their respective books. Threlfall reminds us of several difficulties in its usage.⁶ But whether chemists like the term or not, there is little doubt that it is here to stay. Its usage seems to suggest that there are two structures, the unsolvated and solvated forms, with different crystal structures but that because the systems being considered are different chemical entities the "polymorphism" is not real but "pseudo".⁷ Seddon and others argue, however, that a compound and its solvate must necessarily have different crystal structures, and since they are not chemically identical anyway there is no question of polymorphism, pseudo or otherwise. The literature is confusing, and there are several definitions of the term pseudopolymorph.

There is a certain ambiguity in the term pseudopolymorph. I had mentioned this in my earlier letter,⁸ and there is no getting away from it. But my seemingly casual acceptance of the term arises from scientific considerations. Let me explain. Two factors are paramount in any definition of polymorphism. These are the "sameness" of the chemical substance at the molecular level and the "difference" between the crystalline substances at the supramolecular level. There is little problem with what Seddon writes if one is dealing only with single component crystals because in such crystals the sameness at the molecular level and the differences at the supramolecular level are matters almost beyond dispute. Difficulties arise as soon as the term polymorph is applied to crystals that contain more than one chemical component, and hence the need for the term "false polymorphism" or pseudopolymorphism.

This falseness can arise with respect to the molecular "sameness" or the supramolecular "difference". A serious problem in multicomponent crystals is the identification of the molecular components as being "same", "nearly the same", or "not the same" and in the identification of the crystal structures as being "different", "slightly different", and "not different". Consider a host-guest complex in which the amount of guest is variable and nonstoichiometric. There are many such examples. One may isolate several crystals in which the relative amount of guest varies continuously. Alternatively, one may obtain crystals with a different but chemically or geometrically similar guest. All the crystal structures are "slightly different". All the molecular structures are "nearly the same". It is in this domain of ambiguity that the falseness of polymorphism makes its appearance. Such crystals cannot be called polymorphs because the current definition of polymorphism is too rigid. But the word solvate is too sterile in my opinion. It does not express the difficulties associated in trying to define the structural phenomenon at hand, and it does not show the resemblance of this situation to true polymorphism. The term pseudopolymorphism is not a bad attempt at a definition and reminds me of Churchill's description of democracy.⁹ Still, and to compound the confusion, may I at this point suggest quasipolymorph?

To summarize, the term polymorph should ideally be restricted to crystals of a single compound, when it is trivial to state if two molecules are the same and if two crystal structures are different. In multicomponent crystals, there are too many structural variations and the term polymorphism seems to be unnecessarily restrictive with regard to identity of chemical composition. It is in this context that the virtues of the term pseudopolymorphism need to be considered. This term brings to attention the fact that certain related multicomponent crystals, with slightly different chemical compositions or stoichiometries, have slightly different crystal structures in a manner reminiscent of polymorphism. I agree that indiscriminate application of the term will result in all multicomponent crystals being called pseudopolymorphs, and this we clearly do not want, but it should be sobering for scientists to appreciate that even in our so-called exact sciences there are gray areas in which we should try to comfortably live, without looking for escape routes in surrealistic art.