256 
jection; on the contrary, a classification em- 
bodying the latest scientific conceptions is 
seldom fit for bibliographical work. In the 
Dutch Academy of Sciences ridicule was cast 
upon the decimal system because physiology 
was made a sub-division of medicine. Scien- 
tifically it is absurd ; bibliographically it is the 
only wise course. The literature of the past 
century passes insensibly over into medicine, 
and a system disregarding this historical fact 
would be extremely faulty. All attempts ata 
strictly scientific classification must be personal 
and liable to change. Most zoologists place 
Limulus with the Arachnids ; bibliographically 
this would be folly. Arachnidologists, collect- 
ing the spiders of the various countries of the 
world, have not yet, at least, become so im- 
pressed with this kinship that they seek the seas 
for Limulus; while the malacologists persist- 
ently add Limulus to their lists of captures. 
The bibliographical system should correspond 
with the customs of authors ; it is not intended 
to teach taxonomy. 
The assertion that the decimal system is ine- 
lastic scarcely needs comment. The system 
was first published in 1876, with 1,000 divisons, 
requiring 12 pages of print; to-day by simple 
expansion nearly 50,000 divisions, filling 400 
pages, having been added. For certain sciences 
the expansion has been continued still further. 
Indeed, there are now far more divisions in our 
simple zoological tables than in the entire orig- 
inal work. In certain trials leading up to the 
establishment of the final system used by the 
Concilium Bibliographicum the attempt was 
made to proceed by successive sub-divisions 
down to families and sub-families. In this ex- 
periment as many as a thousand new divisions 
were introduced at a single point in the series; 
it is needless to say that no inconvenience was 
experienced. 
It is a pity to see cautious men of science 
make assertions like this, which have not the 
slightest foundation in fact. They are so plainly 
based upon gross misconception that one might 
well pass them by in silence were it not that 
they are liable to have weight in deciding one 
of the most vital questions now before the scien- 
tific world. 
HERBERT HAVILAND FIELD. 
SCIENCE. 
[N.S. Von. X. No. 243. 
NOTES ON INORGANIC CHEMISTRY. 
THE paper read by Collie and Tickel before 
the Chemical Society (London) on the quadriv- 
alence of oxygen, as shown by the probable con- 
stitution of dimethylpyrone, ‘an oxygen base’ 
has been recently noticed in this column. In this 
paper the authors mention that in 1888 J. F. 
Heyes advocated a similar view to account for 
such peroxides as MnO, and BaO,. In the last 
Chemical News C. T. Kingzett calls attention to 
the fact that in a paper before Section B at the 
Southampton meeting of the British Association, 
in 1882, he reviewed the modes of formation of 
ozone and hydrogen peroxid, arguing for the 
variable valence of oxygen, and adds: ‘‘I am 
not aware that anyone had previously repre- 
sented oxygen as a tetrad.’’ Being present at. 
the Southampton meeting, I remember Mr. 
Kingzett’s paper very well; indeed, I was so 
much impressed with it that I have since 
used the formule O—O'*—O and H,O*°—O 
in my teaching. I recall, however, that 
after the session one of the members re- 
marked to me: ‘‘ Kingzett is right, but there is. 
nothing new in it; I have been teaching that. 
for a number of years.’’ It has long seemed 
strange to me that the idea of the variability 
of oxygen’s valence has had so few advocates, 
especially when its position in the periodic sys- 
tem is considered. 
In a recent number of the Archiv der Phar- 
macie a new method of detecting arsenic in fab- 
ries is given by O. Rossler. A small piece of the 
goods is burned in the upper part of a Bunsen 
flame in a fine platinum spiral, and the arseni- 
ous oxid formed collected on the outside of a. 
porcelain dish filled with cold water. The de- 
posit, which is hardly visible, is moistened with 
silver nitrate. On subsequent fuming with am- 
monia the yellow precipitate of silver arsenite 
becomes visible, and then disappears by solution 
in more ammonia. No data are given as to the 
delicacy of the reaction, but it must be vastly 
inferior to Reinsch’s test, except for such com- 
pounds of arsenic as are wholly insoluble in hy- 
drochlorie acid. In the case of the sulfids of 
arsenic Roéssler’s test might have a consider- 
able value, as the quantity of arsenic present 
in such a yellow pigment is not small. 
