354 The N.Z. Journal of Science and Technology. [Dec. 
may be divided into three categories. The first consists of the proceedings, 
transactions, or journals of learned and professional societies, and the 
reports and bulletins of semi-public institutions and of Government Depart¬ 
ments. By far the greater number of publications of these kinds may be 
procured by gift or by exchange. The second consists of proprietary maga¬ 
zines, issued in general by publishing firms, and obtainable only by purchase. 
The majority of trade journals are included here. The third consists of 
books, also issued by publishing firms, and obtainable only by purchase. 
The first and second categories are serial publications, with one or more 
volumes appearing in each year. The majority of them involve a library 
in a liability for binding, which at New Zealand pre-war prices may be 
estimated at about 5s. per volume. 
Some idea of the extent of the output may be gained from the 
International Catalogue of Scientific Literature , which gives an authors’ 
list and a subject index of all papers published in some seventeen sciences 
since the beginning of this century. The index itself now numbers over 
230 octavo volumes. The number of separate scientific serial publications 
listed up to 1904 exceeded 5,500, and of these many issue more than one 
volume per annum. Of these serials the New Zealand Institute received 
219 in 1917, and it is very improbable that 400 were received in all the 
libraries of New Zealand put together. In the subject of chemistry alone 
the last volume received, dated January, 1919, shows that from 1900 to 
1918 as many as 116,330 separate papers on chemical subjects had been 
listed and indexed. 
It must not be thought that text-books and compilations can supply the 
deficiency of a library in serial publications containing the papers which 
give the results of original research. Text-books treat of only a limited 
number of subjects, and cover.a very small proportion of existing scientific 
knowledge. Except for subjects with a vogue, such as radio-activity, the 
average text-book is several years behind date at the time of its appearance. 
The information supplied is not only often so meagre as to be exasperating, 
but it is only too frequently misleading. Errors once introduced are liable 
to be carried on to subsequent editions and to be copied into other text¬ 
books. The experienced worker uses text-books with caution and relies 
on original papers wherever possible. 
It is not meant to imply by the above-quoted figures that for a library 
to be reasonably efficient for New Zealand investigators it must be complete 
and must include every scientific publication issued. A small proportion 
of them are written in such languages as Russian, Czech, Roumanian, and 
Japanese, and would be of little, if any, use to New-Zealanders. The 
majority are in English, German, French, and Italian ; but even of these 
the larger part consists of the transactions of small local scientific societies, 
dealing only with subjects of local interest, and these would be rarely, if 
ever, consulted in a New Zealand library. 
The figures do show, however, the danger of dividing the available funds 
for library purposes equally between the four chief centres. Such a course 
means an inefficient library in each centre and no efficient library in the 
whole Dominion. Each centre must have a scientific and technological 
library, and it must include the more important serials and text-books, 
but the main effort must be concentrated on a central library in Wel¬ 
lington, and the needs of the whole Dominion must be supplied by a 
liberally interpreted system of co-operation between this library and the 
libraries in the other centres. I propose now to outline the means by 
