NavList:

   

Reply

Frank Reed wrote:

>The accuracy was presumably a few minutes of arc, but I have
>never checked the  observations myself. Anyone?
>
E. B. Knobel investigated the accuracy of Ulugh Beg's star catalogue in
the introduction to his edition of it. After grouping the stars into
northern, zodiacal and southern ones, he shows a systematic bias in
longitude of 18', 14' and 4' respectively. Then he subdivides each group
into 18 sub-groups, each 20 deg of longitude wide. The corresponding
errors in longitude fluctuate about a quarter of a degree around those
means, occasionally a little more. Therefore, some longitudes are more
than half a degree off the truth.  The errors in latitude are of a
similar magnitude. From what I see, they may be just marginally better
than those of Ptolemy/Hipparchus.

The catalogue contains over 100 stars that had not been observed but
whose longitude or latitude was taken over from Ptolemy's catalogue by
applying a (slightly wrong) adjustment for precession, or by referencing
them to an other star. These are not included in the above statistics.

The established latitude of the observatory was not bad. It was too low
by 1 1/2' or 2', depending on which sources you choose.

The star places are tabulated in degrees and minutes. C.H.F. Peters
remarked that the latitudes are all integer multiples of 3', whereas the
longitudes end in 1' mod 3 (i.e. have a remainder of 1' after dividing
by 3). Knobel sees this (particularly the latitudes) as a "clear
indication" that the instrument was graduated in steps of 3'. It's not
so clear to me. It would be a clear indication if the latitudes would
have been observed directly, say, on an armilla. But assuming that the
meridian instrument was used, any odd latitude can result from the
required reduction, whatever the graduation may have been. So there is a
riddle here and I don't know whether it has been solved. Similar for the
longitudes. They must be the result of some final constant adjustment,
but what for? Precession, perhaps? And, again, how did they get rounded
to a multiple of 3' in the first place?

Herbert Prinz

   

Reply