BY E. L. PIESSE, B.SC, LL.B. 



77 



Party- LUt System in Three-party Contest in 1} -member 



Constituency. 



(Size of Larp;est Party, 53^ °/o.) 



48. The tables indicate (as is otherwise apparent) that 

 an odd number of seats for a constituency is to be pre- 

 ferred to an even number, and that all the methods are 

 liable to give somewhat irregular results as the strength 

 of a party changes. 



49. The D'Hondt method has not the advantage that 

 the reasons for it are obvious ; and its merits are still much 

 discussed on the Continent of Europe. (^^) 



50. The rule-of-three method has the advantage of sim- 

 plicity. The condition on which it is based (see the first 

 method, § § 26-35) does not seem to be as just a condition 

 as that of the second method (see § § 36-42), but the cases 

 investigated for a thrje-party contest show that the results 

 are much the same from the two conditions, and indeed 

 that the representation of the largest p&rty by the rule-of- 

 three method is not liable to the curious changes which 

 occur with the second method. For a three-party contest,. 

 then, the rule-of-three method seems to be satisfactory ; 

 but if the number of seats for the constituency is even, it 

 is desirable to add the further condition that a party whose 

 strength is between 50 per cent, and 60 per cent, of all 

 the votes shall have one more than half the number of 

 seats. 



('^) Mr. Humphreys {Proportional Meprtsentation, p. 202), refers to 

 ^the following works, which are not accessible to me in Hobart : — 



Examen Critique dis Divert PraceiUs de Repartition, Proportionnelle 

 en Matiere Ulectorale, by M. E. MacquSiTt, in llevue Scientijiqne, 2Sih. 

 October, 1905. 



La Representation Proportionnelle et les Partis Politiques, by M. P. G. 

 La Chesnais. 



La Vraie Representation Proportionnelle, by Vf. Gaston Moch. 



