Note. In some of the figures and tables accompanying this paper the following abbreviations for parts of the Doric order are used:

¹ Vitr. iv. 3.4.

² BSA lxx (1975) 68–9, 71–2.

³ Bacon, F. H.et al., Investigations at Assos 1881–3 (1902–1921), 113 fig. 1.Google Scholar

⁴ If expressed in terms of a foot of 0·294 m., the capital height is 14 dactyls (to the nearest millimetre); the abacus height is 5 dactyls (to the nearest millimetre), that is 14/3 calculated to the nearest dactyl. The division between annulets and neck then splits the remaining 9 dactyls.

⁵ e.g. Délos viii 250 fig. 119 pl. 27; xxvii 21 fig. 14.

⁶ Penrose, F. C., Principles of Athenian Architecture (2nd edn., 1888) 48–50 pl. 19.Google Scholar

⁷ Hesperia Suppl. v (1941) 122–4; Travlos, J., Pictorial Dictionary of Ancient Athens (1971) 88–9.Google Scholar

⁸ e.g. de la Messelière, Coste in BCH lxvi–lxvii (1942–1943) 53–66Google Scholar; lxxxvii (1963) 639–52. Amandry, in Hesperia xxi (1952) 257–9Google Scholar; Roux, , L'Architecture de l'Argolide (1961) 410–11Google Scholar; Bouras, B., Ἡ Ἀναστήλωσις τῆς στοᾶς τῆς Βραυρῶνος (1967) 149–53Google Scholar; Fouilles de Delphes ii, Michaud, J. P., Le temple en calcaire (1977) 131–5.Google ScholarVallois, R., Architecture héllénique et héllénistique à Délos ii (1966) 143–62.Google Scholar

⁹ A distinction was drawn in BSA lxx (1975) 64–5 between the governing factor, which sets the approximate size an element should have, with some variation allowed on either side of the theoretical value, and the defining factor, which controls the precise dimensions within the allowable range and which may be quite different from the governing factor. It is the governing factor which we are here concerned with.

¹⁰ Note that this is not the same dimension as the ‘saillie d'échine’ used by many of the scholars mentioned in n. 8; capital projection includes the projection of neck and annulets, which is excluded from saillie d'échine.

¹¹ I am most grateful to F. A. Cooper for supplying me with accurate measurements of the capitals of the temple of Apollo at Bassai.

¹² See p. 94 below.

¹³ I am grateful to the University of Edinburgh for providing me with computer facilities for this purpose, and it is a pleasure to record my thanks to the many people who helped my first steps in the world of computers, particularly to J. H. Ottaway and E. Renshaw of the University of Edinburgh, to D. T. Muxworthy, D. D. M. Ogilvie and other members of the staff of the Edinburgh Regional Computing Centre, and to K. I. Macdonald of Nuffield College, Oxford.

¹⁴ BSA lxx (1975) 94–7, Coulton, J. J., Greek Architects at Work (1977) 105.Google Scholar

¹⁵ Dinsmoor, W. B., Architecture of Ancient Greece (1950) 89 and n. 1Google Scholar, associates the order in question with the small southern temple (which he identifies as an earlier temple of Nemesis), and dates both to 487 B.C. Bergquist, B., The Archaic Greek Temenos (1967) 42–3Google Scholar, dissociates the order from the building, identifying the latter as a temple of Themis, and attributing the order to a late sixth-century temple of Nemesis beneath the classical one. See also Boersma, J. S., Athenian Building Policy from 561/0 to 405/4. B.C. (1970) 35, 77–8.Google Scholar For convenience the capital is here referred to as belonging to the temple of ‘Themis’.

¹⁶ This is hard to reconcile with Dinsmoor's view that the temple, begun as Ionic, was only converted to the Doric order in the late fourth century B.C. (Dinsmoor, W. B., The Architecture of Ancient Greece (1950) 184, 221Google Scholar), a view followed by Gruben, G., Die Tempel der Griechen (2nd edn., 1976) 149Google Scholar, and, more doubtfully, by Tomlinson, R. A., Greek Sanctuaries (1976) 75.Google Scholar The fourth- and third-century inscriptions record very little mason's work, and it seems that the building had progressed much further, already in the Doric order, during the fifth century (cf. Délos xii, F. Courby, Les temples d'Apollon (1931) 98–106).

¹⁷ Hesperia xxiii (1954) 40, xxxvii (1968) 44. The transition between Groups 4/5 and 6 is further discussed below, pp. 100–1.

¹⁸ Welter, G., Troitzen und Kalaureia (1941) 45–7Google Scholar, Coulton, J. J., The Architectural Development of the Greek Stoa (1976) 45.Google Scholar

¹⁹ Roux, G., L'Architecture de l'Argolide (1961) 404–5.Google Scholar

²⁰ The transition between Groups 6 and 8 is further discussed below, p. 101–2.

²¹ Vitr. iv. 3.4.

²² Vitruvius (iv. 3.4) gives first place to the abacus height; the designer of the Assos capital discussed above (p. 81 and n. 4) seems to have divided off the abacus first; and the unfinished Attic capitals (p. 82 above) already have the abacus defined at a stage where the annulets could hardly have been rendered.

²³ Vitr. i. 2.4.

²⁴ Hesperia ix (1940) 22.

²⁵ Vitr. iii. 3.1, 3.10.

²⁶ Vitr. iii. 3.12, iii. 5.7, 5.8; Coulton, J. J., Greek Architects at Work (1977) 86–8.Google Scholar

²⁷ These discrepancies arise from the fact that in capital height and abacus height the Bassai capitals come closer to the proportions of Group 4/5, in echinus height to those of Group 8.

²⁸ These assessments are derived from, and can be checked against, the histograms Figs. 10–38.