1. The term is derived from ‘chronos’, time, and ‘desmos’, bond, even if ‘desmos’ is here taken to denote something more like ‘bound’ or ‘band’ than like ‘bond’. Unless a more appropriate and succinct term can be proposed, the word chronodesm may be used with poetic (if not scientific) license. Names may also be appropriate for different kinds of chronodesms. The term ‘monodesm’ may indicate that all data used to derive a chronodesm were obtained at a) a single timepoint or b) during a single timespan shorter than the period of a particular rhythm under consideration. Thus, a monodesm would be a conventional reference interval applicable to a single timepoint or timespan. The term ‘rhythmodesm’ may refer generally to a chronodesm based on any satisfactory mathematical model for a given rhythm, applicable to all stages of the rhythm. If a single cosine curve is an appropriate model, one may use a more specific term: ‘cosinordesm’. If a statistically-significant model with normally-distributed residuals cannot be found, with or without data transformation, one may resort to constructing a ‘merodesm’, in which reference intervals are computed separately at each of several different timepoints or stages of a given rhythm, i.e., as a set of monodesms. The terms may be eventually qualified as to whether it applies to an individual-idiochronodesm — or to a population — panchronodesm. The term may be further qualified as to the model used.
2. The subscript capital i (I) denotes a time used for prediction, which may or may not coincide with an actual sampling time denoted by a small i (i).
3. H.W. Lilliefors, J. Am. Stat. Ass. June 1967, p. 399.
4. W.J. Dixon and F.J. Massey, Introduction to Statistical Analysis, 3rd ed. McGraw Hill, 1969.
5. Documenta Geigy Scientific Tables, 7th ed. Ed. K. Diem and C. Lentner. J.R. Geigy, 1970.