xxxix 



front thin plates of the latter, cast on as large a scale as possible, and 

 closely fitting them at the edges. From what has been said of the pro-' 

 perties of this alloy, it may be judged how much mechanical skill the pro- 

 cess required ; but he succeeded in making a three-feet speculum, described 

 in the Transactions for 1840. It was very light and strong, and defined 

 perfectly, except that with large stars the diffraction at the joints produced 

 four minute rays. For those below the fifth magnitude, and of course for 

 fainter objects, this defect is insensible ; and Lord Rosse believed that in 

 this way only would it be possible to attain the extreme limit of telescopic 

 vision. The construction is stronger than the solid speculum with a 

 third of its weight ; and the cellular distribution of the greater part of 

 the mass enables it to assume the temperature of the atmosphere far 

 more rapidly. During these experiments he was led to an important im- 

 provement in the casting of speculum-metal. When the melted alloy is 

 poured into a mould the surfaces of it in contact with the mould harden, 

 while the interior of the mass is still fluid ; this cools in its turn, and in 

 contracting exerts a powerful drag on the outer crust which gives way, and 

 the whole is shivered. Besides, if the solidification is gradual the mass 

 assumes a crystalline character, which gives it, when polished, a mottled 

 surface. Lord Rosse met these difficulties by forming the bottom of his 

 mould of iron, the sides of sand, and by leaving the top open : the metal in 

 contact with the iron congealed almost in an instant, the sides of the spe- 

 culum more slowly, as the sand is a worse conductor than the iron, the 

 upper part remaining fluid longest ; so that the contraction occurred chiefly 

 at the back of the speculum, where it did no harm, while the front pre- 

 sented a layer of uniform and comparatively tough material. He found, 

 however, that this layer is a little more liable to tarnish than that which 

 is cooled slowly. The process is quite different from the chilling of cast 

 iron, with which it has sometimes been confounded. All this was well ; but 

 now another difficulty was found. Copper in fusion absorbs a large quan- 

 tity of oxygen, much of which it gives out on becoming solid, and speculum- 

 metal appears to possess the same property. Owing to this it is always 

 full of microscopic pores, and the escape of the gas when the alloy is cooled 

 by contact with the iron, as it cannot rise through the viscid film which is 

 formed, probably causes the bubbles and cavities which are found even 

 when no air is entangled in the pouring. To give this gas a free escape 

 was the obvious remedy, and that was effected by making the bottom of 

 the mould of hoop iron, placed on edge and packed so closely that it 

 retained the metal, but was pervious to gases. The plan was so successful 

 that in 1840 he had finished a solid three-feet speculum, and had satisfied 

 himself that even a six-feet was quite practicable. 



He soon made the attempt, and succeeded in April 1842, in obtaining 

 a perfect cast which, after being partly ground, was broken by the care- 

 lessness of a workman: two other failures are mentioned in his third paper; 

 one of which, however, actually gave a disk more than seven feet diameter^ 



