Sdrrut and some Allied Mechanisms. 809 



§ 5. After deriving the mechanisms (b) and (c) from the 

 class (a), I was interested to try to discover any references to 

 them. A hrief account of what I have been able to find out 

 may be set down here : and may perhaps be the means of 

 eliciting further information. 



Among the collection of models left by Professor Robert 

 Willis occurs a specimen of class (a). It departs slightly 

 from generality (see fig. 4) in that the two hinges connecting 

 the pieces here called RAT intersect each other : a non- 

 essential peculiarity. There is also a specimen which would 

 belong to class (/>) if the hinge XY were made virtual instead 

 of actual. Of class (c) there appears to be no specimen. It 

 was in searching*, however, through the voluminous MS. 

 lecture-notes of the Professor, that, though finding no mention 

 of either (a) or (b), I first came upon a reference to the 

 article of Sarrut ; together with references to a quotation of 

 his result in two French text-books f of slightly later date, 

 and a rough sketch, marked with dimensions, of a simple 

 form of the mechanism. It seems quite likely that a model 

 was made, but that it was subsequently dislocated and the parts 

 adapted to other uses ; a frequent custom (so Mr. J. Willis 

 Clark tells me) with the Professor. In any case the parallel 

 motion failed to become known ; and twice at least (not to 

 count the present article) it has been reinvented and repub- 

 lished apparently with but little more success. In 1880 a 

 patent (Specification 5492, Dec. 30) was taken out by H. M. 

 Brunei : and in 1891 it was again invented bv Professor 

 Archibald Barr (Proc. Phil. Soc. Glasgow, March 18, 1891) t 

 V^r to this day it appears to remain practically unknown to 

 mechanicians. Its obvious merits and its long neglect of 

 fifty years seem therefore to be worth insisting upon with 

 some emphasis. Compared indeed with any of its later rivals 

 in the presentment of rectilinear motion, it has obviously two 

 points of great superiority. In the first place, there is an 

 entire absence of any special restrictive metrical relations 

 t<> be satisfied by the dimensions of the parts as a condition 

 of the movement. In the second place, the connecting pieces 

 are bnt four in number, with six hinges ; whereas rectilinear 

 motion of a piece seems not to have been obtained other\vise§ 



* By kind permission obtained at the Engineering Laboratory, 

 Cambridge. 



f Laboulaye, Traits de Cin&matiqiie, 1854, p. 634; Girault, Trans- 

 formation da Mouoement) 1858, p. 267. 



X I have to thank J' ' . G. Baily for referring me to Professor 



Barr: the latter, in his turn, had heard of Brunei's patent from Sir 

 Frederick Bramwell, to whom Brunei had once shown it. 

 Kempe, Hart, and Darboux. 



