422 Transactions of the Royal Society of South Africa. 
contained in the fundamental catalogues of stars on which the almanac 
data are based. Were the effects of these personal discordances absolutely 
uniform for any observer, their resulting effect on either of the operations 
just referred to could be reduced to zero — in the former case by dupli- 
cating the programme and interchanging the observers at the two stations 
whose difference of longitude is under investigation — in the latter by 
relying only on differential observations in which each observer's work is 
combined only with his own determinations of clock error. 
There exists, however, a further complication, the discovery of which 
resulted from the comparison of such differential observations with 
differential observations of a different character, not dependent on time 
observations but on accurate micrometric measurements made either on 
the sky itself or on photographic representations of it. There was a 
general tendency for the apparent places of the fainter stars in relation 
to the brighter stars, as determined by the transit observers, to lag behind 
their positions as determined by other methods. That this tendency could 
be traced to inaccuracy in the meridian observations was established by 
covering the object-glass of the telescope with wire-gauze screens, by 
means of which the apparent brightness of the brighter stars could be 
reduced to correspond with that of the fainter stars. When this precaution 
was adopted the anomaly practically disappeared, and the method afforded 
a means of accurately estimating its amount. This was found to be 
similar in kind for all observers, but to vary considerably in amount from 
one observer to another. 
To obviate difficulties resulting from these "personal" discordances 
an entirely new method of transit observing has been devised by Messrs. 
Eepsold, of Hamburg. In place of the fixed reticule of webs in the field of 
the telescope a single moving web is used which the observer can move at 
will by turning of a graduated drum-head. As this drum-head is turned 
a series of "make" and "break" signals is transmitted electrically to 
the recording tape of a chronograph side by side with similar signals 
registering the beats of the clock. Each "make" and each "break" 
corresponds with a definite position of the rotating drum-head, or, what 
corresponds thereto, of the travelling wire. The operation of observing 
then consists in turning the drum-head by hand so that the star under 
observation is continually bisected by the travelling web ; the clock times 
at which it reaches definite positions, these definite positions no longer, 
however, being marked by fixed webs, are then automatically registered 
on the chronograph. 
The method has been extensively used in field operations for deter- 
mining terrestrial longitudes with satisfactory results, and is already 
adopted in several of the leading observatories. 
The method may, perhaps, be compared to that of a gunlayer on board 
