II — III] Introduction xix 



Therefore bj' formula (i) p. xiii : 



•58 X '42 

 z, = •398,6233 - •oS [1793] + ^ \-^^^^ 



= •39S,6233"j 



- 1040 > = -398,5241. 



+ 48j , 



Or, we might proceed as follows: for the Poll-men ^ (1 — a) = '4817, hence 

 a = •0366. But from Table III, p. 9, which gives z for « : 



a = ^03, z =-398,6603 

 a = -04, z = -398,4408 

 61 = -66, A,= -2194 



A,= 



627. 

 66 X -34 



[627] 



= •398,5225. 



Hence by formula (i) : 



z, = -398,6603 - -66 [2194] + 



= -398,6603 



- 1448 



+ 70 



We conclude therefore that z would be correct to five figures with second dififer- 

 euces, and that for four figures, first differences from either Table II or Table III 

 will suffice. 



If we use formula (ii) p. xiii — Everitt's formula — we find from Table II : 

 z, = -398,6233 - -58 [1793] + "^^ >1^6636 ^^^^^ ^ ■4:2 x^-8236 ^^^^^ 



= -398,6233 



-1040 



+ 13 



+ 5 



and from Table III : 



z^ = -398,6603 - 66 [2194] + 



= -398,66031 



-1448 



+ 37r = -398>5223. 



+ 31 1 



Working with formula (iii), Table II gives us z^ = -398,5242 and Table III 

 Zi = -398,5225 with second differences. We shall not therefore without higher 

 differences get from any of our formulae closer than -398,522 with a possible error 



c2 



= -398,5211, 



•66 X -5644 

 6 



[627] + 



-34 X -8844 

 6 



[627] 



