^{page 87 note 1} e.g. Koldewey, and Puchstein, , Die griechische Tempeln in Unter-italien und Sicilien (1899), p. 132 (cf. n. 3)Google Scholar; Dinsmoor, W. B., Architecture of Ancient Greece (1950), p. 112Google Scholar; Dunbabin, T., The Western Greeks (1948), p. 143.Google Scholar

^{page 87 note 2} Pace, B., Sicilia antica, ii (1935)Google Scholar, 236 ff., suggests this in contradiction to all foregoing discussions. The latest edition of the guide to Sicily published by Touring Club Italiano (1959) repeats it as the accepted view, and so do von Matt, L., Das antike Sizilien, p. 172Google Scholar, and Bernábo Brea, L., Musei e monumenti in Sicilia (1958) p. 114.Google Scholar

^{page 87 note 3} Hittorff, J. I. and Zanth, L., Architecture de la Sicile; Recueil des monuments de Ségeste et de Sélinonte (1870) (Text and atlas). The investigation of the site appears to have been conducted some time before 1830. The blocks are described as well cut (Text, p. 41), and the plan seems otherwise to show very accurately the surviving state of the monument. (These blocks are very tentatively sketched in by Koldewey and Puchstein, PI. xix; it is suggested, briefly, that they may be remnants of a cella, p. 133.)Google Scholar

^{page 88 note 1} Thuc. 6. 6. 2.

^{page 88 note 2} Dunbabin, , op. cit., pp. 335 ff.Google Scholar

^{page 88 note 3} I.G. i¹. 19. See also S.E.G. x. 7, and Tod, Greek Historical Inscriptions, no. 31; Tod dates the alliance to 454–3 B.C.Google Scholar

^{page 88 note 4} Thuc. 3. 86 mentions Athens's intervention in eastern Sicily, but says nothing about the connexion with Segesta.

^{page 88 note 5} Dunbabin, loc. cit.

^{page 88 note 6} Thuc. 6. 11. 7.

^{page 88 note 7} Greek architecture was frequently imitated in strictly non-Greek states. But such monuments were always complete replicas of a Hellenic building; the Nereid monument at Xanthos, for example, built at more or less the same time as the Segestan structure, consists of a complete Ionic tetrastyle temple. See Dinsmoor, , op. cit., p. 256.Google Scholar

^{page 89 note 1} The unfluted and undressed temple of Nemesis at Rhamnous provides a good example of the observation of this rule.

^{page 89 note 2} Cf. Dinsmoor's remarks on methods of construction, op. cit., pp. 169 ff.

^{page 89 note 3} Epidauros I (I.G. iv². 1, 102). This is the from Epidauros, which I hope to republish.

^{page 89 note 4} Baunack, , Aus Epidauros (1890), p. 62Google Scholar, and Vallois, R. (B.C.H. xxxvi [1912], 219), suggest that the peristasis-foundations should have been accounted for somewhere, but they do not go so far as to restore the inscription to comply with their suggestion.Google Scholar

^{page 90 note 1} It is clear that, in the lacuna A 8–9 of I.G. iv². 103, must have stood the accounts for the construction of the peristasis, i.e. before the surviving accounts for the cellawalls. The record of accounts for the Artemis temple, I.G. iv². 1, 106, also demonstrate this order of construction.

^{page 90 note 2} Dinsmoor, , ‘Observations on the Hephaisteion’, Hesp., Suppl. v (1941), 30 ff.Google Scholar

^{page 90 note 3} Dinsmoor also quotes the temple of Aphaia on Aigina as a certain example of this rule; the presence of three columns built up in drums on the north side of the otherwise monolithic peristasis is explained as the final filling-up of a gap which had been left for the moving in of blocks required for die interior. This may well be the reason for the difference in construction. And no doubt a wide gap would have been necessary for moving in epistyle blocks. But Pace, op. cit. p. 237, is surely incorrect in suggesting that the (? comparatively) narrow space between the columns at Segesta would have been an impediment to moving in material for die cella, and diat this is a valid objection to the idea that the Segestans meant to complete the building.Google Scholar

^{page 91 note 1} If one rejects Hittorff and Zanth's evidence for cella-foundations at Segesta, a difficulty arises: one then has to assume that the Segestans, while following the general rules of construction, yet completed the peristasis before even thinking of touching the interior of the temple. The only known parallel to this procedure would be that of the curious fifth-century Telesterion at Thorikos; here, only the peristasis was erected, and there is no sign of foundation-cuttings inside it.

^{page 91 note 2} Even so, it looks as if work on the super-structure of the cella did not start for some time after the peristasis was under way. Three contracts, for quarrying and carting stone for the cella, stand in between that for the peristasis and that for the actual building of the cella, and it may be that the contract to build the cella was taken up several months after the contract to build the peristasis.

^{page 93 note 1} Construction of the peristasis at Epidauros cost about 13,000 dr. It would have contained perhaps one-third of the work involved in the Segestan peristasis. But, even so, we cannot assume that the cost of the Segestan peristasis would have been at least three times greater. The transport of material would have been cheaper; and a concerted civic effort (under skilled direction) would probably have come cheaper at Segesta than the employing of independent contractors at Epidauros.

^{page 93 note 2} Thuc. 6. 46, cf. Diod. 12. 82. 3 ff.