362 
INTERNATIONAL ASTRONOMICAL UNION Proc. N. A. S. 
of the reader, are all too evident to admit of question. The problem is 
how can these results be best attained. 
1. Observatory bulletins. — There should be as few bulletins and as little 
published in them as circumstances will permit. It is better to publish 
in standard journals and purchase reprints. 
2. References. — A nearer approach to standardization of references to 
journals, etc., than now exists should be adopted — librarian's method 
perhaps. 
3. Standard kinds of work. — So far as practicable all results obtained in 
any one standard line of work should be presented in the same form. 
4. Summaries. — Most articles should end with concise and accurate 
summaries. 
5. Catalogues. — Catalogues of scientific literature should give concise 
abstracts. 
6. Economy in typesetting. — Manuscript should be kept as free as prac- 
ticable from all letters, symbols, and combinations that worry the printer 
or annoy the reader. Greek letters, for instance, are expensive luxuries 
— a dime per letter being the estimated added cost of publication. Simi- 
larly, subscripts, superscripts, inferior bars, superior bars, vertical frac- 
tions, etc., should be avoided as far as practicable. 
The following recent suggestions by the London Mathematical Society 
are excellent: 
Instead of Always print 
Suggestions 
V2, Vl3 
V2 or 2^, 1/V2 or 2~^, Vl3 or 
13^ 
yJax^ -\- 2bx + c 
yl{ax^ H- 2bx + c) or {ax^ -f 2bx 
^{a/b) or (a/b) 
V-1 
t or t 
n.n + l.n + 2 
\ n + 1 , |2w, 2"|^ 
nin + l){n + 2) 
(n -h 1)!, {2n)l 2".w! 
X, y, r, B 
a a -\- b a -\- b -\- c 
2' 3 ' 4 
x' y y', r', 6" (by preference) 
la, \{a + 6), \{a -\- b -\- c) 
a -]r b a 
c b c b^ 
(a + b)/c, a/{b + c), a/b + c 
