226 
NATURE 
[JANUARY 7, 1897 
considered definitive for all future time”; but Dr. Gill 
succeeded. By a happy combination of tact and firm- 
ness he was able to resist all attempts to lower the 
standard of survey, and, in the face of grudging support, 
to carry out a considerable portion of his original scheme. 
Maclear’s arc has still to be extended to Port Nolloth, 
and this extension to be connected with Mr. Bosman’s 
excellent triangulation in Bechuanaland ; the eastern 
end of the latter triangulation has to be connected with 
the geodetic triangles near Kimberley, and Kimberley, 
through the Orange Free State, with the Natal triangu- 
lation. When these operations have been completed, 
let us hope in a not far distant future, they will yield an 
arc of 14° of longitude. 
Dr. Gill’s success thus far leads him to look forward 
and ask whether the progress made in geodetic survey 
in South Africa might not be regarded as the first step 
in a chain of triangulation which shall extend con- 
tinuously to the mouth of the Nile. Further he would, 
by an additional chain of triangles along the coast of the 
Levant, and through the islands of Greece, connect the 
African and Greek triangulations. and so, with Struve’s 
great arc of meridian, measure an arc of meridian 105° 
in amplitude. Even though the completion of a work of 
such magnitude cannot be expected for many years to 
come, Dr. Gill has done good service by pointing out 
the mode of procedure that might be adopted, and the 
desirability of at once undertaking the definite survey of 
Egypt. Proposals for such a survey were made soon 
after the British occupation, but, for financial reasons, 
which would probably be still urged, they were not 
entertained. We trust, however, that the commence- 
ment of a work of such economic value to Egypt, and 
of such geodetic importance, will not be much longer 
delayed. 
Dr. Gill was most fortunate in the officer selected to 
take charge of the field operations. Captain (now Lieut.- 
Colonel) Morris, R.E., C.M.G., made, personally, nearly 
the whole of the astronomical observations, and of the 
measurements of the angles of the triangulation, as well 
as the greater part of the computations. He also 
arranged for the transport of the delicate instruments 
over much rough country, managed distant heliostat 
parties, looked after men, horses and cattle, and, with 
much tact, persuaded the farmers to allow beacons and 
observatories to be erected on their lands. When it is 
remembered that the 18-inch theodolite with its stand 
weighed 1042 lbs., and that it was packed in five boxes, 
of which that containing the body of the instrument 
weighed 4oo lbs., the transport difficulties are apparent. 
The able manner in which Colonel Morris carried out 
his duties may be gathered from his Report, which fills 
a large portion of the Blue-book, and the fact that no 
breakdown occurred at any time throughout the work. 
As Dr. Gill well says: “Colonel Morris’ services have 
been such as very few men have the combined physique 
and capacity to render.” 
The survey was carried ‘out with great care and 
accuracy with the improved instruments of the present 
day, and it takes a high place amongst the geodetic 
_ surveys of the world. The standard of measure was the 
Cape to-foot standard Bar A., the length of which, in 
terms of the international métre, was determined at 
Breteuil by M. Benoit, then Chief Assistant of the Inter- 
national Bureau of Weights and Measures. The con- 
stants of two Tonnelot thermometers, for employment 
in the comparison of the base-apparatus with the 
standard, were also investigated at Breteuil under the 
direction of Dr. Brock. The base-measuring apparatus, 
made by Troughton and Simms, consisted of five 10-foot 
steel bars, “‘so mounted as to be conveniently handled 
and aligned, and provided with means for determining 
the instantaneous temperature-of each bar.” The steel 
bars were compared with the standard by means of an 
NO. 1419, VOL. 55 | 
instrument, the Comparateur, designed by Dr. Gill, of 
which a full description, with plates, is given. 
Three base-lines were measured, each by a different 
method. The Natal base was measured throughout its 
whole length, 10,800°45 feet. The line was divided into 
three sections of about 3600 feet, and each section 
was measured backwards and forwards in successive 
horizontal 50-foot lengths. The verification of the base 
showed a close agreement (a difference of ‘ooo2 feet in 
the one case, and of ‘0027 feet in the other) between the 
two outer sections as derived, respectively, by direct 
measurement, and by triangulation from the central 
section. As this appeared to show that accuracy could 
be obtained without the measurement of long base-lines, 
it was decided in the case of the Port Elizabeth base to 
measure only about 5690, feet, and afterwards to prolong 
it both ways to a total length of 17,058'49 feet. The 
measured base was divided into eight sections of 700 
feet, each of which was measured forward in ascending, 
and backward in descending temperatures, and the 
measurements were, as far as possible, completed the 
same day. In the Kimberley base 6000 feet were 
measured, and the prolongation to 14,760°3 feet was all 
in one direction. Here the measured base was divided 
into twelve sections, so as to ensure the completion of 
the forward and backward measurements in the same 
day. The probable errors in the lengths of the three 
bases are: Natal + o'14o inches ; Port Elizabeth + 0’072 
inches ; Kimberley + o'081 inches. The probable error 
of the Zwaartland base, measured by Maclear with the 
Colby compensating apparatus used for the measurement 
of the Lough Foyle base, is + 2°362 inches. 
For the triangulation two instruments were used: an 
18-inch theodolite by Troughton and Simms, and a 
1o-inch theodolite by Repsold of Hamburg. The larger 
theodolite, though an admirable instrument, was incon- 
venient from its great weight and bulk ; and, in spite 
of every care, the pivots, being made of comparatively 
soft material, became slightly deformed by wear. The 
Repsold theodolite was much more portable than the 
18-inch, being less than a quarter its weight; it had 
hardened steel pivots which stood seven years’ hard 
service, and it could be used both as a transit instrument 
and as a zenith telescope. It was consequently exclu- 
sively used, after its introduction, for astronomical 
observations, and the measurement of horizontal angles ; 
and it seems to be an excellent instrument for geodetic 
work in a country where difficulties of transport may be 
expected. The probable error of an observed angle with 
the 18-inch is +0’-49, and with the 10-inch +033. 
These results, especially the latter, compare well with 
those of the best geodetic surveys of other countries. 
Those obtained with the Repsold are only inferior to 
the results of the survey of Saxony (+ 023), and of the 
U.S. Coast Survey—San Francisco and Salt Lake series 
(+025). The interesting table of probable errors of, 
measurement of a single angle in the best surveys given 
by Dr. Gill, shows how far the Ordnance Survey (prob- 
able error + 119) has been left behind by more recent 
surveys, and how desirable it is, from a scientific point 
of view and in the interests of geodesy, that the angles 
of the principal triangulation should be remeasured with 
the more perfect instruments of the present day. 
The measured bases of the South African survey were 
connected by chains of triangles; and other stations, 
including those of Maclear’s survey, were adjusted to 
the system of the geodetic circuit. A good account of 
the circuit solution is given by Colonel Morris in his 
Report. The observations of vertical angles were 
generally taken in the afternoon, when the effects of 
vertical refraction were most uniform, and the heights 
were adjusted on the same principle as the horizontal 
angles. The results were satisfactory. The difference 
between the levelled height of Zwaartkop Mountain 
