392 
ME. F. C. PEXEOSE OX THE OEIEXTATIOX OF GREEK TEMPLES. 
result was the discovery of an error I had made in working out the example, which 
showed that, the orientation angle should have been 271° 2' instead of 269° 12'. 
With this alteratio 2 i, and taking as usual the line of the axis, and with the same 
star, the approximate date becomes 550 b.c. instead of 815. If in this case the 
northern limit of the intercolumniation had been used instead of the axis, it would 
have made the date still earlier. 
Selimts [a newly discovered Temrple named in Kalloway and Puchsteix’s Work 
Me gar on of Demeter''). See Plan. 
This small temple, which lies apart from the two great groups of temples, and near 
the right bank of the little stream (the Selinus), must have been of great 
sanctity, judging from the great number of votive offerings which have been found 
near it. Its plan is remarkable, and would he difficult to explain, except on the 
astro]lomical theory of its orientation. 
It is clear from the angle of its orientation that, if intended for a solar temple, it 
could only have been so at the summer solstice. The eastern axis has the amplitude 
of + 30° 11'. That of the sunrise in that latitude and at the presumed date, here¬ 
after given, and on a level site, would have been -j- 30° 35', but the eastern horizon 
in this case is not level, and when first shining into the temple the sun would have 
had the amplitude of 28° 16'. To account for the presumed error made in setting out 
the work, let it be assumed that the normal amplitude at the solstice of 30^° had 
been ascertained elsewhere, and that at some other time of year the lines of the 
temple had been laid down on the actual site to that angle (for the meridian could be 
ascertained very nearly at any time), and that some jorogress had been made with the 
work, when the solstice came round and the error was noticed. Instead of taking the 
work down and iDeginning again, other means were taken to meet the difficulty; what 
these Avere is very clearly pointed out by the plan. 
It Avould have been observed that the sun’s first ray entering centrally the eastern 
door Avould haA^'e fallen considerably to the iiorth of the niche AAdiich seems to haA'e 
been proAuded for the statue, which niche is centrally placed betAveen the flank walls. 
To meet this difficulty, a narroAver naos AA^as constructed,, of AAdiich the foundations 
remain, hugging the north AA^all of the temple ; the centre of AA’hich, marked also by a 
foundation-stone for the statue, Avould haA^e receiA^ed the ray in the desired manner. 
What further marks the incident is that the southernmost angle of the Propylaca, 
AAdiich gives access to the temple area, is exactly kept clear of the line requmed 
for the sunrise. The altar, AAdiich stood nearly in front of the eastern door, could 
not have interfered Avith it, being itself Ioav, and standing on rather loAver ground 
than the temple. 
The AA^arning star AA"ould, indeed, have to be seen oA^er the roof of the Propylma, 
but there is sufficient difference betAA^een the leA^el of the ground at the tAvo sites to 
