May 22, 1902] 
NA TORE 
93 
with a total of about r00 memoirs and papers, while the Poly- 
technics have contributed about a score. In concluding his 
report, the Principal remarked: ‘‘It is time that London 
should realise that it is not the want of men, or a dearth of 
intellectual effort, which has hindered the University of London 
from taking its place as a great centre of teaching and research. 
Our needs are organisation, which shall make the results of the 
work of the teachers, their assistants and students more fruitful 
and better known as results of which London may be proud, 
and funds to supply them with the materials for their work.” 
SCIENTIFIC SERIALS. 
American Journal of Science, May.—Notes on living Cycads, 
by G. R. Wieland. A study of Zamza floridana. Particular 
attention is drawn to the presence on one of the cones of a 
pinnule of normal form and structure which had grown out 
from beneath the outer hexagonal tip of one of the upper 
abortive sporophylls. As in a similar example described by 
Sir W. T. Thiselton-Dyer, this structure is regarded as a re- 
version, exhibiting evolutionary stages which may be found in 
fossilised forms. To speak of these. growths as ‘* monstrous 
cones” is regarded as almost misleading.—On crystals of 
Croconite from Tasmania, by R. G. Van Name.—Notes on 
unusual minerals from the Pacific States, by R. W. Turner. 
Among the phosphates found were pyromorphite, apatite and 
monazite, the latter occurring in abundance in the Idaho basin. 
—On the use of the stereographic projection for geographical 
maps and sailing charts, by S. L. Penfield. A continuation of 
previous papers on the same subject.—Note on the application 
of the phase rule to the fusing points of copper, silver and gold, 
by T. W. Richards. It has been found by Holborn and Day 
that gold gives a very constant melting point, copper two con- 
stant points at 1065° and 1084° C., whilst silver gives no fixed 
point. It is shown that all these results could have been de- 
duced by the application of the phase rule.—The initiative 
action of iodine and other oxidisers in the hydrolysis of starch 
and dextrins, by F. E. Hale.—Note on the possibility of a 
colloidal state of gases, by C. Barus.—Some glacial remains 
near Woodstock, Connecticut, by J. W. Eggleston. 
American Journal of Mathemattcs, vol. xxiv. No. 2, April. 
—L. E. Dickson, on the canonical form of a linear homogeneous 
transformation in an arbitrary field of rationality. Ina previous 
paper (4. /. xxii. p. 121) the author obtained a reduction to a 
canonical form for transformation in a Galois field ; it is here 
proved that the same process applies when the field is arbitrary. 
—H. B. Newson, a new theory of collineations and their Lie 
groups. A geometrical theory of collineation in the plane, 
independent of Lie’s analytical method of transformation-groups. 
—L. P. Eisenhart, infinitesimal deformation of surfaces. <A 
discussion of the transformation «2’=x+ex, y/=y+en, 
2’ =2+ €2, with dxdx, + dydy, + dzdz,=0, and e a small 
constant, of which the square is neglected. 
SOCIETIES AND ACADEMIES, 
LONDON. 
Royal Society, March 6.—‘‘On the Spark Discharge from 
Metallic Poles in Water.” By Sir Norman Lockyer, K.C.B., 
F.R.S. 
In this paper various modifications produced in the spectra 
of metals by alterations of the conditions under which the 
substances are volatilised are discussed and new observa- 
tions made at the Solar Physics Observatory are described. 
The investigation was undertaken partly in consequence of a 
suggestion put forward by Dr. Wilsing, of Potsdam, to the effect 
that certain conditions, viz. the production of spark spectra in 
liquids, gave rise to the formation of structural peculiarities in 
the constituent lines which are characteristic of the spectra of 
new stars. 
One of the chief characteristics of the spectra of Nove is the 
occurrence of a series of double lines, each consisting of a bright 
and a dark component, the latter being always situated on the 
violet or more refrangible side of the bright line and in contact 
with it. The usual interpretation of this appearance has been 
to consider the composite spectrum produced by two bodies in 
relative motion, but the necessary velocity is greatly in excess of 
NO. 1699, VOL. 66] 
other known cosmical motions. Recent experiments dealing 
with the spectra of elements under pressure having shown that 
by this means the wave-lengths of the lines are altered, Dr. 
Wilsing suggested that if the pressure were sufficiently great, dis- 
placements might be obtained of equal magnitude to those 
observed in the case of new stars. As the direct application of 
high pressures is attended with difficulties, he utilised the fact 
that exceedingly high tensions are produced when electric 
sparks are discharged in liquids. 
Using an induction coil, with jar and air break in the secondary 
circuit, a brilliant discharge is produced in water, giving a very 
intense continuous spectrum crossed by faint metallic lines. In 
this way Dr. Wilsing obtained the spectra of iron, nickel, 
platinum, copper, tin, zinc, cadmium, lead and silver, and from 
the examination of the photographs he arrived at the conclusion 
that displacements of lines and double lines occurred which were 
in every way similar to those in the spectra of Nova Aurigze, 
and that therefore, in all probability, pressure is the cause of 
the duplication and broadening of the lines in the spectra of 
new stars. 
On examining the first few spectra obtained under these 
special conditions, the appearances presented were so suggestive 
of many of the well-known effects of reversal that a further 
inquiry was advisable. It has long been known that in 
ordinary arc spectra many instances occur in which the absorp- 
tion line is asymmetrical with respect to the emission line ; and 
reference is made in the paper to communications by the author 
to the Royal Society more than a quarter of a century ago 
describing these peculiarities in certain silver and rubidium 
lines. 
The experiments at the Solar Physics Observatory were made 
first with the large Spottiswoode coil, capable of giving a 42-inch 
spark in air, this being intensified by the insertion of a large 
glass-plate-condenser in the secondary circuit, so that the 
sparks obtained were about 3 mm. long in air and o°5 mm. 
in water. 
Later a 10-inch coil was used with a smaller condenser in 
circuit, and about the same sparking conditions. The photo- 
graphs of the spectrum were taken on a large scale by means of 
a 6-inch Rowland concave grating of 21°5 feet radius, with 
14,438 lines to the inch. The first-order spectrum was 
employed, arranged to photograph the region from A 3800 to 
A 4800, occupying a length of 18 inches on the plate, Distilled 
water was used in all cases. 
Of the metals examined (iron, silver, lead, copper, zinc and 
magnesium) only iron, magnesium and zinc showed reversals, 
and those of zinc were extremely weak. In all cases the lines 
of the spectrum of the spark in water are much broader than. 
the corresponding lines in the spectrum of the air-spark. From 
an examination of several plates of different intensity, however, 
it appeared that the broadening was, for the most part, of 
similar nature to that observed in the arc spectrum in air when 
an excess of material is introduced between the poles. 
When the cases of non-symmetrical absorption were considered, 
it was noted that very different appearances were presented 
according to the exposure of the spectrum. For example, in 
the best exposed plate of iron, the line at A 4260°64 is well 
reversed in the water-spark, with the part of the emission line 
towards the red several times stronger than the portion on the 
violet side of the absorption. An even diminution of the whole 
composite line, as shown by photographs of less exposure, 
results in the persistence of the less refrangible portion only of 
the emission line, which alone would suggest the presence of a 
line greatly displaced towards the red with regard to the original, 
spark line in air. Several of the iron lines show the inter- 
mediate stage, where the violet component is on the verge of 
visibility, and in these cases the appearance is suggestive of a 
bright line with a dark companion on its more. refrangible 
border. It is important to note, however, that in these cases 
the absorption line is usually normal with the position of the 
original line, the bright component being displaced towards 
the red. 
In the water-spark spectrum of copper it is only with diffi- 
culty that any existing line spectrum can be distinguished from 
the intense continuous emission, and the few lines so recorded 
present the appearance of broad bands, displaced towards the 
red. The probability of their being produced in the manner 
suggested, however, is rendered feasible by the fact that, although 
no actual absorption is visible, their more refrangible edges are 
fairly sharply defined, while the other edges are quite diffuse. 
