to the 21ethod of Least Squares. 209 



1864 WooLHOUSE. 'On Interpolation, Snmmation, and the Ad- 

 justment of Numerical Tables.' Lond. Assurance Mag., Vol. XI, 

 pp. 61-88, 300-y32; Vol. XII, pp. 136-175. 



1865 BRtjNNOw. ' The Method of Least Squares.' Spherical As- 

 tronomy (First English from the second German edition, London 

 and New York, Svo), pp. 40-60. 



A very elementally sketch of the Method. 



1865 Gooss. '' Begrundung der Methode der Meinsten Quadrate.'' 

 Kreutznach, Svo, pp. 32. 



A doctor's thesis. Contains a deduction of the law y ■=- ce ^ 

 from the axioms that the curve is symmetrical, that it has the axis of 

 X for an asymtote, that the equation must be a simple one, etc. The 

 discussion is not very satisfactory. 



1865 Tait. 'On the Law of Frequency of Error.' Trans. Soc. 



Edmh., Vol. XXIV, 7 pp. 



The principle of the investigation is that an error arising from any 

 source may be compared to the deviation from the most probable re- 

 sult of the number of white or black balls obtained by a great num- 

 ber of drawings from a bag containing equal numbers of white and 

 black balls. The idea and the algebraic work is nearly the same as 

 Quetelet's investigation of 1846. See 1872 Glaisher. 



1865 ToDHUNTER. 'A Flistory of the Mathematical Theory of 

 Probability from the time of Pascal to that of Laplace.' Cam- 

 bridge and London, 8vo, pp. xvi, 624. 



This work is invaluable to all students of the Theory of Probabil- 

 ity and I have to acknowledge my great indebtedness to it in prepar- 

 ing the early part of thifc list. None but those who have undertaken 

 such historical researches can form an idea of the immense amount 

 of labor which must have been done in preparing a Avork like this of 



TODHUNTEK. 



Todhunter's analyses of the memoirs of Lagrange and Laplace 

 are full and clear, and his commentary on Laplace's proof of the 

 Method of Least Squares greatly simplities the tedious investigations 

 of the Th'eorie analytique des Probabilites. An account of Gauss's 

 proof of 1809 is not given. 



1866 Borsch. 'Ueber die mittlern Fehler der Resultate aus tri"-o- 

 nometrischen Messungen.' Archio. Math. u. Phys., Vol XLVI, pp. 

 40-44. 



2x^ being the sum of the squares of the residual errors and n the 

 number of direct observations, the mean error has been taken as 

 Tkans. Conn. Acad., Yol. IV. 27 Oct., 1877. 



