274 
(chart plates proper) to be reproduced in facsimile. In 
June, 1900, the state of affairs was as follows :—15,000 of 
the 22,000 catalogue plates had been ¢aken, but only 4000 
had been measured ; and the measurement is of course 
by far the most serious part of the work. Of the 22,000 
chart plates required, less than 4ooo had been taken, and 
only a small portion of these had been reproduced and 
published. So that the fraction of the whole programme 
accomplished in a dozen years can certainly not be put 
higher than one-fifth. 
Does this mean, then, that it will take sixty years to 
finish the whole? It is earnestly to be hoped that 
this would not be a legitimate inference, and fortunately 
there are good sound reasons why it should not be. The 
years immediately succeeding 1887 were naturally devoted 
to experimental work, of which a large amount has been 
necessary. This was foreseen at the outset ; witness, for 
instance, the words of the veteran Otto Struve in his 
opening address :— 
“En effet, ’Astronomie pratique posséde aujourd’hui, 
dans la Photographie, un instrument de la plus haute 
valeur et qui, probablement avec le temps, facilitera 
énormément nos études épineuses. Mais restons sobres 
dans nos prévisions. Pour le moment, nous ne devons 
regarder la Photographie que comme un instrument trés 
précieux, sais dont Pétude reste encore a compléter.” 
But it will probably be agreed that the amount of work 
necessary to “complete the study” has exceeded expec- 
tation. 
Beyond the preliminary experiments which might have 
been foreseen by an individual worker, much time has 
been spent in a well-meant endeavour to secure uni- 
formity in the work, which has, after all, not been very 
successful. Thus a large part of one year was lost in 
attempts to devise an obscuring screen which should 
diminish the light received from the stars in a known 
ratio, and ultimately secure uniformity in the limiting 
brightness (or rather faintness) of the stars charted ; but 
this attempt was at last abandoned in favour of the 
simpler method of fixing a definite time of exposure, which 
might have been adopted from the first. Or going further 
back in the history, it must be remembered that although a 
standard pattern of telescope was adopted in 1887, it took 
a considerable time, not only to make the eighteen 
instruments required, but for the makers to find out how 
to make them. Thus it would be fair to estimate that in 
1900 the work had been in actual progress, not for a 
dozen years, but for less than half that period ; so we 
need not fear that the completion of the work is still half 
a century off. Nevertheless, he would be sanguine who 
should reduce this prospective limit below twenty years, 
unless some very drastic measure is adopted in the near 
future. Some of the cooperating observatories are well 
advanced with their work, but others are far behind. In 
1900 there were actually three which had not started at 
all, and these have been struck off the list and replaced 
by three new ones. We have good reason for antici- 
pating energetic action from these new comers, but it 
must be remembered that they start a dozen years at 
least behind their colleagues. 
This great delay in the execution of the work has been 
prominently mentioned because it demands most serious 
attention if the original scheme is to be carried out in 
any real manner. Even without the addition of the Eros 
work there was sufficient cause for anxiety; with that 
important and unforeseen addition there is reason for 
alarm. It is to be hoped that the dangers may be 
realised and obviated within the next few years. 
But when we turn to the contemplation of what has 
been accomplished, there is good reason for satisfaction. 
To take first the series of catalogue plates, with short 
exposures of a few minutes only. Each observatory has 
to take about 1200 of these, and the area of the 
NO. 1707, VOL. 66] 
NATURE 
[JuLy 17, 1902 
sky covered by each is a square of two degrees in 
the side, so that sixteen full moons arranged in solid 
square formation would just about cover this area. 
On each plate there are some 300 or 400 star- 
images on the average ; but this is an average from 
which the deviations are large. A plate exposed near 
the Milky Way, even for a few minutes only, shows 
thousands of stars, whereas if the telescope be pointed to 
a region distant from the Milky Way, the number may 
fall below 100. Taking the average as 350, there are on 
the 1200 plates which form the share of one observatory 
some 400,000 star-images ; and it is the business of that 
observatory, after taking the plates, to measure carefully 
the relative positions of all these images and publish the 
results. Moreover, it has been found advisable to make 
these measures at least twice over, so that we may put 
the total number at something like a million. It will 
readily be conceded that this is a gigantic piece of work 
for a single observatory to carry out, and it is a great 
thing to be able to say that some of the observatories are 
already in sight of its accomplishment. Others, as has 
been admitted, have not yet commenced the work, but 
they will enter upon it with all the advantages of follow- 
ing an example already set, and we may consider that 
the greatest difficulties have been overcome. 
This portion of the work affords another reason for 
satisfaction. Mention has been made of some pre- 
liminary experimental work which produced no positive 
result, but other such investigations have had more 
fortunate ‘issues, especially the research on the best 
method of measuring the plates. In 1887 there were at 
least three different methods which might be adopted, 
and corresponding to each of these there was a choice of 
patterns for the instrument to carry it out. The proper 
method for measuring stellar photographs has now been 
practically settled, and though there is diversity of 
opinion as to the best actual instrument, the relative 
advantages of the different forms are becoming tolerably 
well known. It will be realised how definite an advance 
has here been made when it is remembered that an 
eminent astronomer, in reviewing the possibilities in 
1887, dismissed the method which has since been 
universally adopted as obviously inferior to the others 
and not worthy of consideration. The test of experience 
had, in fact, not been applied, and the result of its 
application may be regarded as a valuable scientific 
asset. 
Let us turn now to the other set of plates, the chart 
plates as they are called, similar in every way to the 
catalogue plates, except that they are exposed to the 
sky for a much longer time (forty minutes at least, in- 
stead of three or six), and hence contain thousands of 
stars instead of hundreds. It is proposed that these 
plates shall be reproduced on paper by some process 
which depends on the automatic action of light only, and 
is thus free from the imperfections incidental to human 
agency. The exact process has not been formally 
specified, and it is open to any observatory to circulate 
ordinary contact prints, for instance, if such can be made 
without losing too many of the fainter star-images. Up 
to the present time, however, the only reproductions of 
chart plates which have been published are in helio- 
gravure. The French observatories (Paris, Algiers, 
Toulouse, Bordeaux) and the Observatory of San 
Fernando, in Spain, have produced and circulated most 
beautiful enlargements (twice the linear dimensions) of 
some of their chart plates made by heliogravure, and 
there are many reasons why we may hope that their 
example will be universally followed. To begin with, 
the charts are really beautiful to look at—as might be 
expected from the French, they have produced some- 
thing zesthetically satisfactory. Secondly—a matter of 
infinitely more importance astronomically—the charts 
