of Edinburgh , Session 1885 - 86 . 
787 
believe there is but one opinion as to the very great accuracy and 
trustworthiness of Dr Sang’s tables. The methods which he used 
ensured the detection of errors, whether of computation or of tran- 
scription, with almost absolute certainty. Such results, I need 
hardly say, could not be attained without the expenditure of labour, 
perseverance, and unremitting attention, in a degree exceeding that 
required by almost any other human undertaking, hto considera- 
tions, save zeal for the advancement of science and a benevolent 
desire to lighten the labour of future computers, could have induced 
Dr Sang to undertake such a gigantic task, or have sustained him 
through the wearisome mass of mechanical detail which overlaid the 
more interesting parts of his occupation. 
I may conclude with a remark on the subject of Dr Sang’s last 
communication to the Society on this subject, — the paper on 
decimal subdivisions of the circle. Although the author has in a 
manner pledged himself to the system of decimal division of the 
quadrant in the elaborate table of sines which he has constructed 
on that basis, he is willing to allow that there may be some good 
in other methods, and, indeed, indicates that he does not view with 
disfavour the rival proposition of the decimal subdivision of the 
degree. 
So far as the business of astronomical computation is concerned, 
it really makes no difference whether the circle is to be divided into 
1000 equal parts, into 360 degrees, or into 24 hours. But the sexa- 
gesimal division of the hour and the degree into minutes and seconds 
is a real disadvantage, increasing at once the labour of reduction, 
and the risk of making mistakes in readings. If the degree be 
selected as the unit of angular measure, and no doubt it is a very 
convenient unit, undoubtedly the fractional parts of this unit ought 
to be expressed as decimals of a degree. Apparently the only reason 
for the non-adoption of this most desirable reform is the want of 
convenient logarithmic tables adapted to the decimal subdivision of 
the degree. The quantities corresponding to tenths and hundredths 
of a degree (being 6' and 36" respectively) can be taken directly from 
existing tables, but the intermediate numbers to thousandths of a 
degree, or 3" -6, must be got by interpolation. It is to be hoped that 
Dr Sang’s suggestions on this subject will not be lost sight of, 
because the publication of decimal trigonometrial tables, based on 
