536 



NATURE 



[June 15, 191 1 



much of the work of a character which had hitherto be(;n 

 performed by reference to the sky could be effectively and 

 rapidly carried out through the medium of photographic 

 representations thereof. The series of photographs secured 

 at the Cape formed a promising field for research, and he 

 voluntarily undert<x>k the laborious task of the measure- 

 ment of these plates and the preparation of a catalogue, 

 which has since been issued among the publications of the 

 Cape Observatory, and to-day is in constant use by 

 astronomers concerned with the southern hemisphere. 

 Besides many incidental results of high interest, the cata- 

 logue forms the basis for statistical investigations of stellar 

 distribution, &c. 



The success of the project, even in the early stages of 

 its execution, pointed to the possibilities of a still more 

 extended use of the photographic methods, and gave a 

 direct stimulus to the study of the details by which its 

 efficiency could he increased. The surveys of Argelander 

 and Schonfeld, completed by the " Cape Photographic 

 Durchmusterung " (as the above catalogue is entitled), and 

 independently by a visual " Durchmusterung " emanating 

 from the observatory at Cordoba, under the direction of 

 Dr. Thome, could but be regarded as preliminary pioneer 

 surveys ; but, stupendous as the task seemed, the possi- 

 bility was recognised by Sir David Gill of utilising the 

 photographic method for the production of a complete map 

 of the heavens, which should be on a sufficiently large 

 scale and sufficiently precise in its detail to meet all the 

 requirements of modern astronomy of precision. He 

 succeeded in securing the interest of astronomers generally, 

 and in particular of Admiral Mouchez, the then director 

 of the Paris Observatory, at whose instance, with the «up- 

 jx)rt of the Paris Academy of Sciences, an international 

 conference of astronomers was invited to meet in Paris in 

 1887 to discuss the possibility and desirability of carrving 

 out this extended project. While the plan iii outline "met 

 with universal acceptance, much useful discussion took 

 place at this and at subsequent conferences as to the 

 details of its execution, such as the type of instrument to 

 be employed, the scale of the photographs, the duration of 

 exposure which would serve to bring out what should be 

 considered essential detail, &c. The outcome was that an 

 agreement was arrived at by which the execution of the 

 work was partitioned among eighteen cooperating observa- 

 tories, who between them undertook to secure photographs 

 of the whole sky with instruments of similar dimensions 

 and design, and, so far as possible, similar conditions of 

 exposure. Two series of photographs were to be taken, 

 one of long exposure, designed for direct photographic 

 reproduction to form an atlas which should show all stars 

 down to the fourteenth visual magnitude, and a second 

 series, of shorter length of exposure, designed for exact 

 measurement, which would form the basis of a catalogue 

 of precision of all stars to the eleventh magnitude. Minor 

 details were largely left to the discretion of the directors 

 of the cooperating observatories. 



It would take too long to describe in detail the pro- 

 gress of this work from its initiation, and I propose, there- 

 fore, to confine my attention to the share in it which has 

 been taken by the observatorv over which I have the 

 honour to nroside, a share which, partly from the origin 

 of the project and p:irtly from the scarcitv of suitably 

 situated observatories of a sufficiently permanent character 

 in the southern hemisphere able to "offer cooperation, has 

 exceeded that assigned to any other single establishment. 

 The type of instrument agreed upon was the photographic 

 refractor, similar in dimensions to that which has been 

 used by the Brothers Henry in Paris, giving pictures on 

 such a scale that i millimetre on the plate corresponds 

 with a minute of arc on the sky. The plates were to be 

 160 mm. square, so as, apart from the marginal edges, 

 to yield an effective field of views of two degrees square. 

 The area covered by each plate will perhaps be better 

 presented to you by the statement that the length of the 

 edge of the plate corresponds, roughly, with four times 

 the apparent diameter of the sun or nioon, i.e. four suns 

 and four moons photographed side bv side on the same 

 plate would just reach from one edge of the plate to th^ 

 other. On this scale it would require rather more tlinn 

 10,000 plates to cover the whole sky. The programme. 

 however, provided not merely for the photographing of 



NO. 2172, VOL. 86] 



each region once, and once only, but that the various 

 regions covered by the separate photographs should over- 

 lap in such a way that every star would be contained in 

 two of the regions at least. Thus any doubtful featur«'». 

 shown in one plate could be verified by reference !■ 

 second plate, which includes the same area. This x' 

 inent raised the total number of separate regi'""^ 

 photographed to 22,054, ^^ which 1512 were 

 the Cape Observatory. The construction of th 

 and the provision of suitable housing accommodation wa?. 

 sanctioned in 1888, and the telescope arrived at the Cajv- 

 in 1890. About two years were spent in its ei- 

 adjustment, and testing, the delay being largely i 

 the necessity for returning the object-glass for nei 

 alterations and the remodelling of the breech-piece, 

 could not be effected locally. From July, 1902, on 

 the work of taking the photographs was steadily pro- 

 ceeded with until its completion in 1910. At the present 

 time we have stored at the Cape three almost complet< 

 series of photographs of the 1512 regions, besides many 

 duplicates and triplicates, which, though not attaining th< 

 desired standard of perfection, arc still available for refer- 

 ence with regard to many issues that arise. Of thes< 

 plates, one series of long exposure is destined for direct 

 photographic reproduction ; the other two series of shorter 

 duration of exposure for measurement for the purposes of 

 the catalogue. On account of the comparatively insignifi- 

 cant amount of work involved in taking these catalogue 

 plates as compared with the more laborious work of 

 measurement, after its original completion in 1896 it was 

 decided to repeat the whole series of catalogue plates, so 

 as to bring the dates of the photographs nearer to the 

 epoch 1900, and thereby more fully attain the ideal before 

 the International Conference of leaving for a future 

 generation as exact a record as possible of the appearance 

 of the heavens at the beginning of the twentieth century. 



After some early experiments with the view of ascertain- 

 ing the most economical means by which these plan ~ 

 might be measured with the desired accuracy, the definitivf 

 programme of measurement was commenced in 1897, and 

 completed last year. The number of stars measured on a 

 plate varies between 45 and 3431, the total number of 

 star images which have been measured exceeding a 

 million. To illustrate the precision attained, the discord- 

 ances in the position of a star, relatively to the stars 

 immediately surrounding it, as derived independently from 

 the measures made on two overlapping plates, but rarely 

 amount to so much as a second of arc, i.e. apparent 

 angular distance between the two edges of a halfpenny 

 at a distance of about three miles from the observer. 

 Now, however exactly the measurable features of each 

 plate may be determined, it will be clear that the informa- 

 tion thus derived can be of little value without some 

 e.xternal data to enable the identification of the region 

 on the photograph with the corresponding region of the 

 sky from which it is derived. A plan of this rocwn might 

 have been useful to enable you to find your seats to-night 

 in positions corresponding to the numbers on your tickets, 

 but I venture to think a stranger visiting Cape Town for 

 the first time would prefer to have been provided with a 

 map of the town, which would indicate the position of 

 ;*his hall, rather than to have to search the town for a 

 room arranged according to such a plan. So with our 

 photographs, the detailed study of each photograph can 

 be of minor value without some indications enabling us 

 to identify the field shown on the photc^raph, not only 

 in relation to the surrounding fields, but in relation to the 

 sky as a whole. It is certainly a desideratum that these 

 indications sliould be not less precise than those involved 

 in the measures themselves. Of course, if each photo- 

 graph contained some familiar group of stars, such, for 

 instance, as the constellation Orion, the identity of the 

 region could not be easily mistaken ; but the scale of 

 which the plates are taken is such that two such familiar 

 stars are rarely to be found within the limits of a sinf'le 

 plate, while nany plates are devoid of stars to which 

 accurate positions could hav? been previously assigned. 



The necessity his thus arisen of determining with hich 

 accuracy the absolute positions as opposed to th** relativ* 

 positions indicated by the plate itself, or at least two st.T-s 

 contained on each plate. The distance between the 



