The perfection of a discovery

Some useful preliminaries can be extracted from earlier chapters. For example, tests can determine whether the ion or the molecule is the biologically active species (Section 10.7). Whether it can chelate with a metal cation may be suggested by a glance at the structure (Section 11.2) followed by a quick pH test (Section 11.3). Whether the molecule is structurally non-specific can be found by comparing the minimal hypnotic activity with that of an established example such as chloroform, or by similar comparison of partition coefficients (Section 15.0). 

From this point onwards, the paths diverge. One pathway, as used in this book so far, is to pursue the scientific study of the cytological, biochemical, and distributive clues that provided the lead in the first place. The other pathway is the statistical correlation of several physical properties of the lead molecule with its biological actions. It must be emphasized that statistics lies remote from experimental science and can, at best, indicate only a probability. However, these statistical methods are much used, particularly in industry, and it is proposed to give a brief account of correlation analysis (quantitative structure-activity relationships, or QSAR) which embodies them. 

Quite the most popular of these approaches is multiple regression analysis introduced by Hansch in 1968. In this method, the most favoured variables are (i) partition coefficients (P), from the system octanol/water (Section 3.2) (ii) the sigma (a) and rho (p) values from Hammett s Linear Free Energy Equation (Appendix III) and (iii) Taft s steric factorst E ) which are 

It is evident that multifactorial equations cannot fail to give good correlations with experimental data, because the values of the various constants (k) are systematically altered by the computer until they do so, a result which does not necessarily contribute to a solution of the main problem (Topliss and Costello, 1972 Unger and Hansch, 1973). 

In choosing the first compounds to submit to multiple regression analysis, tremendous help can be derived from Craig s sigma-versus-pi scatter diagram (Fig. 16.i). Representative compoimds must be drawn from each of the four quadrants to discover the aln region of maximal activity. 

---

The rapid thrust of discovery has furnished many new and exciting data which have been incorporated in space vacated by material of less current relevance. At the end of Part Two a new chapter has been added on the perfection of a discovery, regardless of whether the latter was chanced upon, or found by the exercise of logic (Chapter 16). The chapter on selectivity through comparative cytology has been largely rewritten with special attention to cell de-differentiation and cancer (Chapter 5). The chapter on pharmacodynamics has been more usefully restructured (Chapter 7). The former, very short chapter on free radicals has been dismantled, and its contents distributed to more relevant chapters. 

As happens so often in science, a new and more precise technique of measurement led to a major discovery. When scientists first used mass spectrometers they found—much to their surprise—that not all the atoms of a single element have the same mass. In a sample of perfectly pure neon, for example, most of the atoms have mass 3.32 X 10-26 kg, which is about 20 times as great as the mass of a hydrogen atom. Some neon atoms, however, are found to be about 22 times as heavy as hydrogen. Others are about 21 times as heavy (Fig. B.6). All three types of atoms have the same atomic number so they are definitely atoms of neon. 

The nomenclature of a science ought to be distinguished for its clearness and simplimty but it is by no means easy to seoure these conditions in a science like chemistry, where the rapid progress of discovery necessitates the continual addition of new and the frequent alteration of old namea The chemical name of a substance should not only identity and indi-vidualise that substance, but it should also express the composition and constitution of the body, if a compound, to which it is applied. The first of these conditions is readily attained but the second is much more difficult to secure, inasmuch as our ideas of the constitution of chemical compounds—the mode in which they are built up as it were— require firequent modification. On this account all attempts to frame a perfectly consistent system of chemical nomenclature have hitherto been only partially successful. 

Marignac 3 also showed that previous methods for the separation of niobium and tantalum were far from perfect, and for the first time he succeeded in preparing pure niobium and tantalum compounds. His methods are still in use to some extent, and his analyses provided the first reliable values for the atomic weights of these elements. It should be stated, too, that Rose s earlier researches, which extended over a period of nearly twenty years, have provided a valuable source of information for the chemistry of niobium and tantalum.4 His calculations and formulae were revised by Rammelsberg 5 in the light of subsequent discoveries. 

The Absolute Temperature Scale. The idea of the absolute zero of temperature was developed as a result of the discovery of the law of Charles and Gay-Lussac the absolute zero would be the temperature at which an ideal gas would have zero volume at any finite pressure. For some years (until 1848) the absolute temperature scale was defined in terms of a gas thermometer the absolute temperature was taken as proportional to the volume of a sample of gas at constant pressure. Since, however, no real gas approaches a perfect gas closely enough at practically useful pressures to permit an accurate gas thermometer... 

---