ALTAZIMUTH SOLAR OBSERVATIONS. 357 
motion of the sun were that indicated by the arrows in 3 Fig. 4, 
the observations might be made thus :— 
O: IV, time alt. az.: reverse: II, time alt. az: RO 
and if four observations were desired, the continuation might with 
advantage be as follows :— 
RO: I, time alt. az.: reverse: III, time alt. az.: RO. 
The Roman figures denote the position in which the sun is to 
be observed, see the figure. 
In the case of observations of a single tangency as 2 Fig. 4," 
the opposite side of the sun would be observed in the second 
instance, 
Reference has already been made, viz., in § 18, to the impossi- 
bility of employing the double tangency method with telescopes 
of high power, and to the fact that for good determinations of 
direction, the leading or following limb must perforce be observed, 
while for good values of the altitude the upper or lower limb must 
be taken whenever the azimuth component of the sun’s motion is 
not extremely small as compared with its motion in altitude, it 
becomes necessary to ascertain with precision both elements for 
the same moment. Although this cannot be done directly it may 
be readily effected by using an electrochronograph. The routine 
of observation may be as follows, say for a morning observation :— 
Observe (i.) upper limb and time: (ii.) leading limb and time: 
(iii.) lower limb and time: 
For a second observation, reverse face of instrument, and then 
Observe (iv.) lower limb and time: (v.) following limb and time: 
(vi.) upper limb and time. 
By spherical geometry, and the theory of refraction and parallax, 
the altitude of the sun’s centre for the moment of the leading or 
following limb observations, may be readily computed. In 
8eneral there will be a slight discrepancy between the computed 
— observed differences of altitude between observations (i.) and 
(iii), and (iv.) and (vi.): the interpolated values of the altitude 
od 1 for time: the motion being supposed to be nearly vertical. 
