298 



HYDROGRAPHICAL SURVEYING [chap. xiii. 



Sets of 

 Observa- 

 tions. 



Observa- 

 tions 

 must be 

 Similar. 



throughout the day ; but we are in this case hmited by the 

 range of the sextant. 



It is difficult to lay doAvn any rule as to what is the smallest 

 rate of motion in altitude we should observe at, as the greatest 

 motion in altitude during the day varies so much with the 

 latitude and declination. We can only say that we should, 

 when we have any choice, not observe beyond an hoiu* when 

 the time of changing 10' of double altitude exceeds 30 seconds. 



Opinions have much differed on the number of consecutive 

 observations that it is best to take to comprise in a set. The 

 only theoretical limit is that the equation of equal altitudes 

 should be practically the same tlu'oughout the set, as the 

 variation in the time required by the sun to traverse the 

 number of minutes of altitude between observations at the 

 beginning and end of the set will not matter, as we do not care 

 whether the mean of the times agrees exactly with the mean 

 of the altitudes. 



It seems well, therefore, to observ^e tolerably long sets, as 

 errors of observation are thereby eliminated. The same result 

 in the end will be attained by a large number of shorter sets ; 

 but the value of each set is much enhanced if composed of a 

 considerable number of observations, and it saves time and 

 trouble in the calculations. 



Too long sets are to be deprecated as wearying to the eye 

 and hand, and the observations will therefore suffer from that 

 cause, especially in hot countries, where the necessity for 

 observing in the full glare of the smi makes it a trying operation. 



We prefer to take eleven observations in a set. Tliis allows 

 the observer to commence his second set, of lower limbs (a.m. 

 observations), at exactly 1° more altitude than his first one, 

 of upper limbs. It does not much matter, as each one has 

 his own plans for these details, and soon falls into a regular 

 method, which is the great thing to prevent mistakes. 



The only point, in fact, of importance is always to observe 

 in the same way. Not only does it .save time and errors, but 

 it is necessary in combining observations, whether for rate or 

 meridian distance, that they be as similar in all respects as we 

 can get them. The whole system is a system of differences, 

 and it is manifest that the result is the better the more like 

 the observations are. It follows from this that the observers 



