43 'Experiments on Wooii^. 



cellular, replete with laiuinsc, and abound in veins of mal* 

 leahle iron. 



ft is probable, had the native of Hindostan tlie means of 

 rendering; his cast steel as fluid as water, it would have oc- 

 curred to him to have run it into moulds, and bv this means 

 have acquired an article uniform in its quality and conve- 

 ijieut for those purposes to which it is applied. 



The hammering, which is evident around the feeder and 

 upon the upper surface in oent-ral, mav thus be accounted 

 for : — When the cake is taken from the pot or crucible, the 

 feeder will most probablv be slightly elevated, and the top 

 of the cake partially covered with small masses of ore and 

 steel iron, which the paucity of the heat had left either InS' 

 perfectlv separated or unfused. These most probablv, to 

 make the product more marketable, arc cut off at a second 

 lieating, and the whole surface hanmicrcd smooth. 



I have observed the same facts and similar appearances 

 in (jperations of a like nature, and can account satisfactorilv 

 for it as follows : 



The fu'st portions of metal that are separated in experi- 

 ments of this nature, contain the largest sliare of the wdiole 

 carbon introduced into the mixture. It follows, (jf course, 

 that an inferior degree of heat will maintain this portion of 

 metal in a state of fluidity, but that a much higher tempe- 

 rature is requisite to reduce the particles of metal, thus for 

 a season robbed of their carbon, and bring them into eon- 

 tact with the portion iirst rendered fluid, to receive their 

 proportion of the steely principle. Where the heat is lan- 

 guid, the descent of the last portions of iron is sluggish, 

 the mass below begins to lose its fluidity, while its disposi- 

 tion for giving out carbon is reduced bv the gradual addition - 

 of more iron. An accumulation takes place of metallic 

 masses of various diameters, rising up for half an inch or 

 more into the glass that covers the metal ; these are neailv 

 ■welded and inserted into each other, and diminish in dia- 

 meter as they go up. The length or even the existence of 

 this feeder or cxcrescenec depends upon the heat in gene- 

 ral, and upon its ten)perature at diflerent periods of the 

 same process. ' If there has l)cen sufficient heat, the surface 

 v.ill be convex, and uniformlv crvstallinej Jiut if the heat 

 has been nrsred, after the feeder has been formed and ati 

 atfinity established between it and ihe^teclified mass below, 

 it will onlv partially di-<appear in the latter, and the head 

 oc part of the upper end of the feeder will be iound suspended 

 in the «ilas> that covers the steel. 



TliC 



