1919.] Thomson.—Libraries of Science and Technology. 
359 
In order that such a scheme might be easily operated a standard form 
of index card, and of library classification and subject index, should be 
adopted. The two latter requirements are discussed fully below. The 
index cards should include not only the author’s name and the title of the 
book, together with the library classification decimal number and the 
subject-index notation, but also full bibliographic details as shown below, 
including the weight for purposes of estimating postage, and an indication 
of the library or libraries in which the book is filed, such indication being 
most conveniently the library accession number. I put forward a suggested 
form and size for the index card for the purpose of inviting suggestions for 
its improvement. For serials a second card of the same size is desirable, 
differing only in the details of author’s name and title. 
Name of Author. 
Suter, Henry 
Library and Subject Classification. 
594-931 
Title. 
Manual of the New Zealand Mollusca 
Page-size. 
21 X 14 cm. 
No. of Pages. 
Figs, and Plates. 
Publisher. 
Place of 
Publication. 
Date. 
Binding. 
Weight. 
lb. oz. 
3 2 
xxiii 
1120 
See Atlas 
Govt. Printer 
Wellington, 
N.Z. 
1913 
Cloth 
Bom. L. 
A 1550 
Gen. Ass. L. 
Turnbull L. 
Well. P.L. 
Viet. Coll. L. Auclc. Inst. L. 
Audi. P.L. 
lude. Univ. Coll. L. 
Cant. Phil. Inst. L. 
Cant. P.L. 
Cant. Coll. L. 
Otago Inst. L. Dun. Cam. L. 
Otago Univ. L. 
Proposed Card for recording Accessions to the Dominion and other Libraries. 
Two cr more copies of each of these cards should be filed for public refer¬ 
ence—one indexed according to author’s name, to enable the inquirer to 
ascertain whether and where any given book is available ; the others indexed 
according to subject, to enable the inquirer to ascertain easily the literature 
on any given subject. In respect to the latter need, however, the subject 
index would be very incomplete, in view of the predominance of serials in 
the various libraries, unless every paper in each volume—and there may be 
as many as a hundred—were separately indexed. This is undoubtedly the 
ideal method of indexing, but its adoption would involve so much additional 
expense that I hesitate to advocate it for fear of imperilling the whole 
scheme. If it is not adopted, there are two ways in which this need may 
be met. One is to procure for each of the four centres complete sets of 
the International Catalogue of Scientific Literature, the annual subscription 
to which is £17 10s. per annum. This catalogue, however, deals only with 
seventeen sciences and neglects many branches of technology. The other 
method is to have a staff of abstractors in the Dominion Library to whom 
