to the Method of Least Sqiuu-es. 177 



1826 [Fourier]. Kecherches statistiques sur la ville de Paris et 

 le dc'paitement de la Seine [par le Compte Chahrol-de-Volvic]. 

 Paris, 4to, Vol. Ill, in 1826, Vol. IV in 1829. 



These volumes contain investigations by Fourier on weights, 

 probable errors, etc. See note by DeMorgan in Loud. Assur. 3[a(/., 

 Vol. XIV, p. 89. 



1826 Ivory. ' On the Method of the Least Squares.' Tillodi's 



Phil. Mag., Vol. LXVIII, pp. 161-165. 



Contains Ivory's third attempted proof of the Method, which is 

 still more absurd than those of 1825. See Ellis's analysis in his 

 paper of 1844. See also below, 1830 Francoeur. 



1827 Rosenberger. ' Ueber die, auf Veranstaltung der fraiizo- 

 sischen Academic, wahrend der Jahre 1736 und 1737 im Schweden 

 vorgenommcne Gradmessung.' Astroii. JV^achr., Vol. VI, col. 1-32. 



An application of a method, communicated to the author by Bes- 

 sel, for the adjustment of geodetic triangulations by the use of 

 Least Squares. See 1831 Hansen. See Ahhandl. oon IBessel (Leip- 

 zig, 1875), Vol. Ill, pp. 16-19. 



1827 . LTeber die Theorie der Zuverlassigkeit der Beobach- 



tungen und Versuche und der von derselben abhangigen Bestim- 

 mungen des Mittels aus gegebenen Zahlen. [Berlin]. 



1828 Bessel. ' Ueber die Bestimmung des Gesetzes einer period- 



ischen Erscheinung.' Astron. JVcichr., Vol. VI, col. 333-348. — Ab- 



handl von Bessel (Leipzig, 1875, 4to), Vol. II, pp. 364-372. 



Proposes the periodic function since so much used for discussing 

 recurring phenomena, and illustrates its application to the determi- 

 nation of empirical formuUe. See 1864 Sohott. 



1828 Gauss. ' Suj)plementuin theoriie combinationis observationum 

 erroribus minimis obnoxia?.' Cornme/it. Sor. (r<">tti//r/e/i, Vol. VI, pj). 

 57-98. — Also Gauss Werke, Vol. IV, pp. 57-93. — French trans. 

 see 1855 Bertrand, 



This memoir discusses a method for the combination of observations 

 when the observed quantities are not expressed as explicit functions 

 of the unknown quantities to be determined ; and when the problem 

 furnishes rigorous equations of condition which the determined values 

 of the unknown quantities must exactly satisfy. The method of cor- 

 relatives for the adjustment of such conditioned observations is given 

 and an algorithm presented for its use. A numerical example involv- 

 ing twenty-four observations subject to thirteen conditions illustrates 

 the use of the formulae. See the Gott. gelehrte A/tzeigen, Sept. 25, 



Tbans. Conn. Acad., Yol. IV. 23 Oct., 1877. 



