234 
NAT ORE 
[JANUARY 3, 1907 
conditions. A later paper by the same author describes a 
pocket spectrophotometer embodying some novel features. 
In the paper on the platinum-point electrolytic detector 
for electrical waves Mr. Austin describes the so-called 
““ barretter ’’ patented by Fessenden, and used by him as 
detector in wireless telegraphy experiments. It consists of 
a cell with electrodes, one a fine platinum point, the second 
a plate, the vessel being filled with an electrolyte giving 
gaseous decomposition products. When an E.M.F. is 
applied to such a cell polarisation ensues, so that scarcely 
any current passes unless the E.M.F. exceeds a certain 
critical value. When electric oscillations pass through the 
cell the resistance is decreased, and the current for the 
moment increases. Conflicting statements have been made 
by various investigators regarding the behaviour of the 
instrument, and the author has therefore subjected it to 
a thorough investigation, employing both ordinary alter- 
nating current waves and also the Hertzian waves from the 
station of the National Electric Signalling Company. The 
chief conclusions of the research are :— 
(1) For the stronger alternating currents used the break- 
ing down in resistance is approximately proportional to the 
square of the alternating current. 
(2) Under favourable conditions and with moderate 
polarisation the detector is equally sensitive to alternating 
currents with the point electrode, anode, or kathode. 
(3) The resistance of the detector for slowly alternating 
currents varied from 20,000 ohms to 400 ohms, according 
to the polarising E.M.F. employed. 
(4) For electrical waves from a distance the detector is 
approximately equally sensitive with the point electrode, 
anode, or kathode, but for waves from a coil in the 
laboratory some cause appears to annul the sensitiveness 
of the kathode-point electrode. 
The next paper, by Prof. Coffin, is a mathematical in- 
vestigation on the influence of frequency on self-inductance, 
and is not capable of useful abstraction. 
Messrs. Guthe and Austin then deal with experiments on 
the magnetic alloys discovered accidentally by Dr. Heusler, 
and previously investigated at the Reichsanstalt and by 
Messrs. Fleming and Hadfield. Curves of permeability and 
inductance were determined for seven different samples, the 
chemical analysis of which is also given. An ingenious 
apparatus of high sensitiveness, quite cheap and easy to 
construct, was designed for study of the magnetic ex- 
pansion of the alloys. This was capable of detecting 
changes in length as small as 5x10-7 mm. The investi- 
gation is not complete, but the relations between the curves 
of magnetisation and magnetostriction and between 
magnetostriction and thermoelectric force are clearly shown. 
The number and variety of the subjects dealt with in 
these two instalments of the official publication of the 
Bureau show that, though only established three years ago, 
it has already begun to make substantial additions to our 
knowledge of physics. J. A. Harker. 
RESEARCHES IN STELLAR PARALLAX. 
OME years ago Dr. Chase, of the Yale University 
Observatory, communicated to the Astronomical and 
Astrophysical Society of America the results of a survey 
which he had carried out, in collaboration with Dr. Elkin, 
in order to detect stellar parallax. The number of stars 
examined was ninety-two, and these were generally selected 
from a list of stars having an annual proper motion of 
more than half a second of arc. Of these ninety-two stars, 
fifteen had a negative parallax, and, presuming that some 
of the smaller positive values were equally untrustworthy, 
some sixty were left which exhibited a real parallax 
amounting to more than o’-05. The scheme was one that 
seemed worthy of further prosecution, since the method 
. 
employed proved adequate for the purpose of recognising. 
the existence of measurable parallax. Consequently, this 
work has been very considerably extended, and the recent 
publication from Yale gives the details of the discussion 
of no less than 163 stars, forming a contribution of the 
1 “ Parallax Investigations on 163 Stars mainly of Jarge Proper Motion.’ 
By Frederick L. Chase, Mason F. Smith and William L. Elkin. Transac- 
tions of the Astronomical Observatory of Yale University, vol. ii. part i. 
Pp. 207. (New Haven: The Observatory, 1906.) 
NO. 1940, VOL. 75] 
highest importance in parallactic inquiry. Some thirteen 
years have been devoted to the completion of this work, 
in which, though Dr. Elkin and Mr. Smith have taken 
part, the heat and burden of the day has been borne by 
Dr. Chase. This will be seen from the following tabular 
statement, which shows both the class of stars examined 
and the distribution of the work among the several 
observers :— 
Chase FhAS® Smith Smith Eypin 
Stars with proper motion over i 
OMA es j ae seal TZ fees) AS) ves PS) Re 
Stars selected by De Ball on 
account of magnitude veo Wise ns ae 
8 Cygni, Algol; Nova Persei.. 5 ... — ... =~ ... — p= 
Red(starsiforcolonur effect...) (6)-0—"..) — 5 
Totalinumber of series ... 139) =. 5 .-. 14) =) 130 eazO 
The inquiry has been based entirely on measures of 
distance made with the heliometer. As a rule, two com- 
parison stars were selected on opposite sides of the star 
the distance of which was to be investigated, and in the 
direction of maximum displacement by parallax. These 
comparison stars were as nearly as possible equidistant 
from the principal star. When the arrangement of the 
stars did not permit this programme to be carried out 
completely, special artifices had to be adopted. On the 
other hand, in the case of stars of particular interest, a 
larger number of comparison stars was selected. The 
precautions which Dr. Elkin found necessary in the course 
of his work on the parallaxes of stars of the first magni- 
tude were applied here, and further reference to these 
details is unnecessary. Finally, the observations, when 
collected, usually give for each star twelve complete obsery- 
ations, consisting of four groups of three nights each, taken 
at those seasons of the year when the parallactic displace- 
ment was at its maximum. 
A suspicion having been aroused that the measures of 
distance between two stars of different colours might need 
an additional correction for refraction, a series of observ- 
ations was made on some strongly coloured red stars taken 
from Kruger’s ‘‘ Catalog der farbigen Sterne.’’ A term 
was introduced into the differential refraction correction of 
the form AB tanz cos(p—q), where p is the position, q the 
parallactic angle, z the zenith distance, and Af the colour 
effect sought. The several values of AB are as follows :— 
Star Colour AB Weight 
Scale WI a 
Kruger 985 6:0 —0°019+0°019 63°6 
3 LOSO. 70 +0'005+0°020 64°4 
SLOT Ol mee weit) Via. s41010007+-O'OL5 » . a NTOSO 
yy LEST Pon co (72S 0,014 01018) ano 
on DIOR: wares ae toy, +0'046+0'017 45'2 
We Bi Ver 745 ccs se —0'003+0°021 55'6 
The authors contend from these figures that the mean 
light of the red star is apparently refracted less than that 
of the comparison stars. Whether this conclusion is 
justified or not, the quantities involved are so small that it 
can be safely asserted that there is no noticeable vitiation 
in the parallax results arising from this cause within the 
probable errors. The remark attributed to Sir David Gill, 
that the tendency of the heliometer observer is to bring 
the similarly coloured parts of the star’s spectra into 
coincidence rather than the brightest parts, seems to gain 
additional support from this investigation. 
Of the 163 stars examined, the parallaxes range from 
—o".13 to +0".20, and the number of negative parallaxes 
is thirty-six. Considering how wide the net has been 
spread to catch any star, the proximity of which might be 
suspected on various grounds, the chance of finding stars 
closer to us than those which have already been examined’ 
grows very slight. The scheme of the stellar universe, so 
far as the few stars nearest to us are concerned, is taking 
fairly definite shape, and the scale that has been adopted 
from measured parallax will probably need no material 
alteration. Such a conclusion is the more warranted, 
because the precision attaching to the mean value of a 
group of results is far greater than that of any individual 
determination. The authors insist upon this point, and, 
