NovEMBEK 29, 1906] 
VL Chie 
113 
cent. CaO. The orthosilicate is readily attacked by water, 
which dissolves out the lime in large quantities; this is 
probably the reason why it is not found as a_ natural 
mineral. 
The three eutectics are :—(1) tridymite+metasilicate, 
37 per cent. CaO, 1417°; (2) metasilicate+orthosilicate, 
54 per cent. CaO, 1430°; (3) orthosilicate+lime, 67} per 
cent. CaO, 2015°. 
Although the melting point of lime is unknown, the 
authors have been able to plot a complete diagram of the 
different equilibria that may occur in this series of com- 
pounds (Fig. 1). The importance of such an achievement 
ean scarcely be overestimated, and there can be little doubt 
that it will play as important a part in the development 
of experimental mineralogy as Roozeboom’s classical 
diagram for the iron-carbon steels has done in modern 
metallurgy. T. M. Lowry. 
CHARACTER AND CAUSE OF SUN-SPOT 
SPECTRA. 
[t is now just forty years since the selective widening 
of Fraunhoferic lines in sun-spot spectra was first 
observed by Sir Norman Lockyer. Since then various 
papers relating to the same subject have been published by 
Sir Norman Lockyer, Prof. Young, and—more recently— 
Dr. Mitchell and Prof. Hale. 
The authors of the present paper state at the outset 
that in considering the chief features of sun-spot spectra, 
three points especially attract attention :— 
(1) The fact that certain lines in the spectrum of a 
given element are strengthened, others are weakened, the 
remainder being unaffected. 
(2) The occurrence of the strengthened lines in the visible 
spectrum only; none appear in the ultra-violet. 
» (3) The relatively great intensity of the continuous back- 
ground of the spot spectra in the less refrangible region. 
From what is known of laboratory spectra taken 
under varying temperature conditions, the following facts 
accrue :— 
(1) That in passing from a high temperature to a lower 
one, certain lines are relatively strengthened, some are un- 
affected, and others are diminished in intensity. 
(2) That such a reduction of temperature is accompanied 
by an increase in the relative intensity of the less re- 
frangible lines, and a shift of the maximum of a con- 
tinuous spectrum towards the red. 
The general correspondence of these two groups of facts 
led the authors to seek for an explanation of the spectrum 
of sun-spots, on the hypothesis that the metallic vapours 
within the spots have a temperature lower than that of 
the photosphere. 
Photographs of spot spectra, made with the Snow tele- 
scope and a Littrow spectrograph of 18 feet (5-5 m.) focal 
length, and showing a great number of affected lines, were 
available for the investigation. The range of spectrum 
covered by these photographs is from D to Hf. Supple- 
mentary photographs of the spectra of recent large spots, 
extending from A in the red to the ultra-violet, have been 
obtained by Mr. Ellerman. 
The laboratory work began with a study of iron and 
other metals in a synchronous rotating arc, designed and 
constructed by Prof. Crew, but as the necessary photo- 
graphs involved undesirably long exposures this was not 
continued. It occurred to Mr. Gale to try the effect of 
varying the current strength in an ordinary r10-volt direct- 
current arc, the difference of potential between the poles 
being kept approximately constant. Photographs were 
taken, with currents of 30 amperes and 2 amperes, of the 
spectra of iron, titanium, vanadium, chromium, manganese, 
calcium, and other metals characteristic of sun-spots. 
As the work progressed, a correspondence was noted 
between the enhanced lines (lines stronger in spark than 
in arc) and those weakened in sun-spots. To get further 
light on this, photographs were taken of the spectra 
1 Abstract of an advance proof from Mount Wils n Solar Observatory of a 
“*Preliminary Paper on the Cause of the Characteristic Phenomena of Sun- 
spot Spectra.” (Contributions from the Solar Observatory, No. 11.) By 
George E. Hale, Walter S Adams and Henrv G. Gale. The paper is to 
be published in a future number of the Astrophysical Journal. 
NO. 1935, VOL. 75 | 
| 
of the same elements in the discharge of a 6o00-watt trans- 
former, giving about 6000 volts at the secondary terminals. 
A condenser was used in the discharge circuit, and the 
potential was increased by an auxiliary air spark in series 
with the observed spark, both being exposed to a strong 
blast of air from an electric fan. Under these conditions 
| the enhanced lines of the spark are well shown. 
The instrument used to obtain the majority of the labor- 
atory photographs was a grating spectroscope in the 
Littrow form of 13 feet (3:96 m.) focal length. The 
Michelson grating has 7oo lines to the millimetre. In 
taking the comparison photographs of strong- and weak- 
are spectra, two of the strong-are spectra, with varying 
exposures, were generally placed on each side of the weak- 
are spectrum. From the four different strong-arc ex- 
posures thus obtained that one was selected which was 
most nearly comparable in general strength with the weak- 
arc spectrum. In some cases the spark spectrum was 
added, adjoining the weak-arc spectrum, with the strong- 
are spectra arranged as before. 
Tables are given in the paper which contain the results 
of a study of the elements titanium, vanadium, iron, 
chromium, and manganese for the region extending from 
the ultra-violet to A 5800. The tables include all the lines 
which are affected prominently, and which, being 
strengthened or weakened in spots, or in spark or weak 
arc as compared with strong arc, are of special importance 
in the investigation. 
In discussing the behaviour of the lines of the elements 
mentioned, the investigation is divided into two parts, 
the relation of the weak are to the strong arc, and of 
the arc to the spark. Two sets of tables are therefore 
given. The first shows the wave-lengths of all the lines 
which are much affected in spots, the amount by which 
they are affected, their behaviour in the weak as compared 
with the strong arc, and in the spark as compared with 
the weak arc. The second set of tables gives a compari- 
son of the intensities of the lines of the same elements 
which are considerably enhanced in the spark, with their 
intensities in the weak arc. The majority of these which 
occur in the less refrangible part of the spectrum are 
weakened in spots, and such lines therefore appear in 
both sets of tables. As, however. most of the strongly 
enhanced spark lines occur in the violet and ultra-violet— 
where the spot lines seem to have the same intensity as 
the Fraunhoferic lines—independent lists of these lines 
have been added, since the evidence afforded by them as 
to the relation of spark to weak arc is extremely important. 
The authors summarise some of the results accruing from 
the investigation as follows :— ; : 
(1) More than 90 per cent. of the lines in the tables 
which are strengthened in sun-spots are found to be 
strengthened in passing from a 30-ampere arc to a 2-ampere 
arc. 
(2) More than 90 per cent. of the lines shown by the 
tables to be weakened in sun-spots are weakened or absent 
in the 2-ampere arc. - q 
(3) More than 90 per cent. of all the enhanced lines in- 
cluded in the tables are weakened or absent in the 2-ampere 
are. ; : 
(4) In a list selected at random of 152 lines which are 
not spot lines, no cases were found of lines strengthened 
in the low-current arc. ‘ 
In discussing the temperature hypothesis as the probable 
explanation of the observed phenomena, some of the points 
made are :— wen 
(1) Waidner and Burgess’s investigation of the tempera- 
ture of the arc showed that the temperature of the crater 
was reduced 70° when the current was reduced from 30 
to 15 amperes. As the relative intensities of the lines 
undergo no material change in passing from go to 15 
amperes, while the change from 30 to 2 amperes Is very 
pronounced, the temperature of the crater is probably con- 
siderably reduced at 2 amperes. Re tan 
(2) Since the enhanced lines of the spark diminish in 
intensity in the 30-ampere arc, and are still further re- 
duced in the 2-ampere arc, no explanation hitherto 
advanced to account for these lines appears adequate in 
the present case, unless it be the explanation based on 
change of temperature. 
