Notes on the Desir/n of Telescope Tubes. 23 



to contend with, as in the bridge girder, all in one plane. The 

 tube must be a girder in at least two different planes. Now, 

 two ordinary girders, intersecting each other at right-angles, 

 would be well adapted, as far as strength and stiffness are 

 concerned, but are optically inadmissible ; and therefore it is 

 necessary to fall back on a prismatic section, each side of the 

 prism being a complete girder. A prism of four sides — a 

 square section — would be strong and stiff, but somewhat 

 unsightly. It has been employed by no less an authority 

 than Warren de la Rue in the reflector which he used for 

 obtaining his celebrated photographs of the moon. I have 

 here a model (Fig. 1), in which I have endeavoured to show 

 what appears to me the most favourable disposition of mate- 

 rial, all things considered. It is hexagonal in section, having 

 booms at the anoies, which toQ:ether contain about half the 

 material of which the tube consists. The booms are united 

 by a series of small, straight, diagonal bars, making an angle 

 of 45° with the booms, this being the mathematically demon- 

 strable angle of economy in such structures. The latticed 

 tube ends in a stiff, hexagonal angle-iron ring, as shown. 

 The salient feature of the model is the size and number of 

 the booms ; and this is a very favourable arrangement in 

 view of stiffness, for, as Bindon B. Stoney has shown in his 

 excellent work on Strains in Girders and Framed Struc- 

 tures, the deformation of a girder due to compression or 

 extension of its booms is a large quantity compared with 

 that due to the compression or extension of the smaller bars 

 uniting the booms. 



In contrast to Fig. 1, let us consider Fig. 2, which is a 

 representation of the actual tube of the gTeat Melbourne 

 Telescope. Here we shall, I think, find a systematic infrac- 

 tion of all the canons above laid down. In a properly 

 designed framed structure all the bars are straight ; in the 

 Melbourne Telescope they are all curved. In a properly 

 designed girder a large proportion of the material is placed 

 in the form of longitudinal booms ; in the Melbourne Tele- 

 scope absolutely none is so employed. The proper angle of 

 economy and efficiency is 45°; in the Melbourne Telescope 

 this angle is nowhere found. 



The action of the various bars of the Melbourne Telescope, 

 when under strain, is rather intricate ; I will, howeverj 

 endeavour to trace it. When the tube is horizontal or 

 inclined, the effect of gravity is to produce a longitudinal 



